Transport and Town Planning: The City in Search of Sustainable Development

By Jean Laterrasse

In a context where climate change urgently requires us to alter our paradigms, this book explores the possibilities of cities that are both more energy efficient and more respectful of the environment.

Based on the observation that urban planning has been detrimentally affected by the compartmentalization of knowledge and practices, this book is conceived as a dialog between transport and urban planning on the one hand, and between engineering and social science on the other. Systemic analysis and a historical approach, integrating the teachings of the last two centuries, constitute at the methodological level the framework in which this dialog unfolds.

Based on examples of good practice, Transport and Town Planning identifies an effective set of levers of action and proposes an original method to guide and accompany urban transition with a large share of the initiative reserved for the actors concerned.

• Language: English
• Publisher: Wiley
• Release date: Jan 23, 2019
• ISBN: 9781119579489

Book preview

Preface

This book is based on the observation that a double division marks the theories and practices implemented in urban manufacturing: division between town planning and transport, followed by division between engineering and social sciences. Although these divisions are particularly significant in France, they do not concern France alone but are found in various forms on the five continents, and they today constitute obstacles to progress with the necessary scale and pace towards sustainable urban development.

Taking the opposite view of these reductive cleavages, this book is built around a dialog between these different fields. This dialog is essential today to meet the challenges posed by our contemporary cities. Systemic analysis, on the one hand, and the historical approach integrating the teachings of the last two centuries of urban history, on the other hand, constitute at the methodological level the framework within which this dialog takes place. Conceptual models are used where necessary, but by always explaining the underlying assumptions and opening the black boxes to allow readers who do not have particular engineering knowledge to understand the origins and scopes of application (including the limits) of the models used. From this perspective, the book integrates the long pedagogical practice that I acquired while directing a master’s program that included third year students of top engineering schools as well as architecture students and urban planners in their final year of study.

As a result, this book is intended for students as well as practitioners, for engineers as well as urban planners or managers of our contemporary cities. Its primary purpose is not to provide general and definitive responses, which would seem particularly illusory and inappropriate in the contexts of the great diversity that we all know, but points on which to reflect and tools to better communicate and jointly find, far from normative paths, ways for the future.

With the above being the objective, this book is fueled by many borrowings and exchanges that I managed to have throughout my professional adventure, and different functions that I had the opportunity to carry out, in higher education and research, but also in operational positions. To remain within the recent period, I would like to thank especially my colleagues from the Laboratoire Ville Mobilité Transport (LVMT) for their contributions. This laboratory, on the campus of the Cité Descartes in Marne-la-Vallée, has the particularity of prospering at the crossroads of three cultures: that of the Ecole des Ponts, that of the French National Institute for Research into Transport and Transport Safety (which later became the Institut Français des Sciences et Technologies sur les Transports, l’Aménagement et les Réseaux (French Institute of Science and Technology on Transport, Planning and Networks)) and finally that of the Université Paris Est and its school of urban planning. I equally extend my gratitude to all colleagues from the Energy Transition Institute Efficacity where I work today, and especially its Director M. Salem-Sermanet, who allowed me to carry out these exchanges in a particularly stimulating environment in tune with the theoretical and operational issues involved in the development of decision support tools adapted to the challenges of the contemporary city.

My special appreciation also goes to the Master’s students of Transport and Planning, common to the Ecole d’Urbanisme of Paris and the Ecole des Ponts. This book largely reproduces the course that they helped me to prepare, through their questions and also by the richness and diversity of their contributions, often drawn from case studies backed by some 20 countries from which they originate. Many names come to mind. Many of them with whom I maintained contact occupy important functions today. I have limited myself to mentioning throughout the text only those whose end-of-course dissertations or theses occupy some parts of this work, with a particular mention for Th.B.Aw and M.El Hadeuf who helped me to build respectively Chapter 3 and Chapter 6, but my gratitude goes far beyond and concerns the 15 or so batches I had the pleasure of accompanying.

A very warm thank you to all. I reserve a special thanks to P. Devoyon, without whom I would not have succeeded in formatting the manuscript, and who provided his assistance with great relevance and kindness.

While I benefited for the drafting of this book from many contributions, and that notwithstanding, I acknowledge the responsibility for the positions adopted herein. I

tried to support them with as rigorous a scientific and ethical approach as possible, deeply inspired by the conviction that urban transition will only be made with all stakeholders of the city, in particular citizens, with the latter being able to intervene at every decision-making stage.

If is up to you to appreciate if this book, as I hope, can contribute to it.

Jean LATERRASSE

November 2018

1

City and Complexity: How to Untangle the Skein?

Reflecting on cities and their future quickly leads to a paradox: for a large majority of us, the city is a familiar object, within which we quite easily identify our roots. However, thinking and managing the city is nonetheless a very complex exercise. How can what sounds simple in common sense be at the same time so complex, not to say inextricable, to practitioners and managers? To support our observation, it seemed useful for us to discuss this paradox.

1.1. Systemic thinking and its historical context

The concept of complexity is not trivial, although it can be concealed, as is the case with the object city, under a deceptive simplicity. Within the meaning understood here, this concept is in fact quite recent.

In the late 19th and early 20th Centuries, the history of ideas was marked by a linear vision of progress. In the field of knowledge, it is illustrated by positivist thought: at the same time that it proclaims the supremacy of experimental sciences, it also claims to reduce them to data only accessible by experience and critically underestimates the role of theories. With regard to society, this vision is based on the conviction that economic development, even unequal, benefits everyone and constitutes a kind of royal road to the wellbeing of humanity. The industrial revolution earmarked the triumph of machinery, which was extolled by positivist thinkers, and found in Taylorism a form of organization which seemed to foster such unlimited economic growth and opened to humanity the path of universal progress. At the same time, we are convinced of the almost inexhaustible nature of natural resources, and of the possibility for the economy to develop at the pace of the ability of our societies to transform fossil resources into energy.

World War I, followed by the Great Depression of the 1930s, undermined this vision. The explanation cannot be limited to the observation of the dissensions stirred up by the distribution of wealth created by economic development at different geographical scales. More deeply, under the impetus of the industrial revolution, human society constructed entities – cities, companies, large administrations or big institutions and States – whose operation raised problems that were increasingly difficult to manage. Where it was initially imagined that the only limits to development included the amounts of energy that could be mobilized, we realize that other limitations could be related to management or organizational problems of major economic, social or political constructions produced by human societies.

At the scientific level, the movement of ideas experienced the same type of questioning, and positivism was contradicted by developments that it had not been able to anticipate. The theory of relativity and wave mechanics in physics, or the discovery of genes in biology and the debate between mechanism and vitalism ¹, revealed a natural world that was profoundly different from that opened by empiricism which was so valued by positivists.

From the resulting controversy emerged the idea that human development, marked until then by a strong coherence between the development of productive forces and the movement of scientific ideas, was confronted with what could be referred to as the challenges of complexity, challenges raised by natural, physical or biological systems as well as by artificial systems designed by human societies.

How can this notion of complexity be formalized, what types of concepts and tools could be imagined to understand it, that is, particularly with regard to artificial systems, to manage or operate these complex objects², and also to design new objects that are better adapted to the needs of our societies? These are the questions that, in the pivotal period of the 1930s–1950s and until today, mobilized a large part of scientific activity, both in the field of exact sciences and engineering sciences as well as social sciences, then in full structuring.

1.2. The system approach

This expression has been somewhat improperly used and to excess. Since we use the expression in this work, we deem it appropriate to clarify the content beforehand.

Implicitly or explicitly, the method of thinking that scientists have been referring to for centuries is to consider that we know an object, regardless of its nature, when we can both identify its elements and describe the essential characteristics of each of these elements. This is the so-called analytical or Cartesian method (because of the particularly clear explanations given by R. Descartes in his Discourse on Method³). However, this method of thought, supposedly universal (all those things which fall under the cognizance of man might very likely be mutually related in the same fashion (…), and there can be nothing so remote that we cannot reach to it, nor so hidden that we cannot discover it wrote Descartes), is inadequate to address the concept of complexity. Indeed, it appears that the complexity of natural or artificial objects is not only due to the nature or the more or less high number of elements which constitute them (we would rather talk of a greater or lesser degree of complication), but also due to the multiplicity of interactions between these elements. The perspective of this systemic method is, from this observation, to mainly focus on the combinatorial relationships that more or less strongly inter-relate the different elements of an allegedly complex object, and we will then say that this object – or set of elements – constitutes a system.

This concept of system can be illustrated from a basic but essential application. In the classical deterministic approach, a direct link is postulated between a cause and its effects: the same cause is supposed to always produce the same effects (principle of causality), and the significance of the effects is proportional to the intensity of the cause (principle of linearity) [DUB 75]. The observation of complex systems – whether natural or artificial – often undermines these principles: in the health field, for example, the effects of an epidemic or air pollution will not be the same on a subject in good health and on a subject whose health has been weakened by other illnesses. We thus observe in many cases that the effects of an action depend on the state of the system to which this action applies, a state defined at a given moment by the relations between the elements of the system (for the human body, the relationships between the organs and vital functions). The butterfly effect of meteorologists could also be mentioned here: a local event may, because of the complexity of the systemic interactions involved in a meteorological phenomenon, have significant consequences on a regional scale.

Similar examples are easily found in the area of the city or transport: a passing shower, which makes the road slippery, if the road network is not too dense, will simply be reflected by a longer travel time for vehicles because of the precautions that users must take to avoid collisions; but if the network is dense, even a minor degradation of traffic conditions can result in generalized embolism. Network effects, which can mitigate or, on the contrary, amplify certain phenomena (for example, the consequences of a road accident), are often effects which are systemic in nature.

This concept of system calls for reflection on another notion, which is important in the field of the city, that of optimization: the optimization of a system, within the meaning understood here, implies not only that we improve each element of the system, but also that we work on the relationships between these elements. In the industry, it is known from experience that significant gains in productivity can be obtained on a process by a better articulation of all the operations that compose it⁴. In team sports, we know in the same way that proper coordination between team members is an essential factor for good performance. Many examples can also be cited in the field of transport: in the rail field, the HST (high-speed or intercity train) is the result of a better coupling between a whole set of techniques involving both the track, traction, and structure of the trains. Hybrid vehicles also provide a good illustration of this approach: better engine performance is obtained through a set of operations in the entire energy management chain (involving electrical assistance) and transmission of this energy to the traction members.

1.3. Analytical and systemic methods are complementary rather than opposed

Analytical and systemic methods are often presented in an antagonistic way. Personally, I have always stood against any kind of religious war. These methods are more complementary than they are opposed.

The analytical method seeks to apprehend an object of study in itself, and for this purpose, to accurately identify and distinguish this object’s constituent elements from one another. The systemic method rather focuses on the behavior (or functioning) of this object and its effects (or its performance). To this end, we will rather focus on the observation of relationships between the elements that constitute this object, and between this object and its environment, without ever considering a priori as acquired, neither its boundary (some elements may belong either to the set considered or to its environment) nor even its structure (we can envisage different combinations with the same elements). Moreover, each object is often itself only an element of a larger whole, whose perception may be essential for a good understanding and therefore for the study of the object considered⁵. We will recognize here the difficulty that may be involved in urban geography with regard to isolating the neighborhood from the city, or the city from its hinterland.

The rigorous description of a system would require taking each element that constitutes it into account including their characteristics, as well as relationships between these elements. The analytical and systemic approaches are each a kind of reaching the limits: the former is adapted to systems with weakly related elements and for linear and reversible development processes (for example, around a point of equilibrium); the latter when interactions between elements are strong and developments are nonlinear and irreversible. The first leads to a local, detailed and static vision (an image that could be provided by a microscope), and the second to a global vision, seeking to dynamically apprehend overall behavior or performance (the image would rather be that produced by a macroscope).

We will show below that the two points of view – microscope and the more global one, macroscope – are both essential for thinking the relationships between city and transport: according to the circumstances, we will use one or the other of these visions, and often both simultaneously.

1.4. Transdisciplinarity of the concept of system and presentation of a typology of complexity

At this stage, we will consider that a system is first of all a conceptual tool, which aims to apprehend the characteristics of complex objects, whether they exist in nature or they are the product of human societies. In fact, the concept of system was used in science long before the concern of modeling complex systems emerged. Addressing this issue led K.E. Boulding to propose a typology⁶ of these systems according to their nature.

He distinguishes nine levels⁷:

– the level of static structures or frameworks: anatomy and geography of the universe, solar system, etc. These objects are systems, within the meaning understood above: they are composed of elements that interact (particularly via the gravitational force);

– simple dynamic systems, an example being the clock. Part of the theoretical structure of physics falls into this category;

– cybernetic systems, characterized by the existence of a control mechanism. Classical example: the thermostat. These systems differ from previous ones in two respects: the transmission and interpretation of information becomes an essential part of the system; in addition, equilibrium is no longer simply provided by characteristic equations, but can be modified within the limits determined from factors outside the system. In this category, we have artificial systems (for example, technical systems) and also natural systems: the homeostatic model in physiology is a system of this type;

– open or self-maintaining systems, an example being the living cell. They mark the boundary between living and non-living. The ability to sustain themselves and reproduce that characterizes these systems depends on their degree of openness⁸. The more the system becomes complex, the more this concept of openness becomes significant, and the more the systemic analysis will become a relevant method of knowledge;

– at a higher level including that of more elaborate biological systems (for example, plants), we will find systems characterized by the emergence of division of labor between cells, with differentiated and mutually dependent parts (roots, leaves, etc.). In this type of system, sense organs and information receivers are poorly developed: as a result, their information reception capacity, and thus their degree of openness, remains limited;

– animal systems: here we witness the emergence of increased mobility, a behavior that can be described as teleological (animals act as if they had an objective). At this level, more specialized information receptors emerge, the brain which enables a processing of this information, and enhancement of the nervous system. The intervention of images between stimuli and animal response generates behaviors that cease to be a constant function of stimuli, with more or less elaborate forms of adaptation⁹ and learning;

– human systems have more elaborate functions that Boulding describes as the transition from self-awareness to self-consciousness. The mental representations that they produce are more complex, and characterized by an ability to turn into oneself: not only human beings know, but also they know that they know (access to language and symbolism). More concretely, they have the capacity not only to adapt to their environment, but to act to adapt this environment to their own needs. Moreover, human beings are different from animals by an increasingly elaborate view of the passage of time;

– social systems, characterized by the behavior of individuals in society: the distinction of this level with the previous one appears somewhat arbitrary, insofar as human beings are in essence social individuals. But at the level of collectivity and its organized expression emerges a complex set of functions and activities: production, exchanges, spatial planning, political functions, etc.

– finally, Boulding distinguishes a ninth level, in which he brings together "transcendental systems, which relate to the mystery of life".

In this typology, each level simultaneously integrates, completes and surpasses the previous one. It will not have escaped the reader that each level corresponds at least in a first approach to different disciplines. This typology, beyond its didactic character, does not presume to constitute a scientific demonstration. It is not exhaustive, in the sense that certain intermediate levels would be necessary for the introduction of ecosystems in this classification or animal species such as social insects (bees, termites, ants, etc.) whose behavior is fascinating to study. Another major drawback is that the transition between physicochemical and living systems does not appear clearly, and that the higher level of complexity stems more from a metaphysical vision than a truly scientific approach.

This simple typology will, beyond its historical interest and despite its imperfections, allow us to introduce several important ideas:

– the first is that social systems appear, if the metaphysical dimension introduced by Boulding is excluded, as the most complex. They refer to the triptych being human/nature/society, etc. However, the systems which we discuss in this book, urban systems, are both technical systems (which have the particularity of being deployed generally over vast territories) and social systems (they are designed and constructed by human societies, and for their activation and use, they always presume the effective intervention of organized human beings). We will often refer to them as sociotechnical systems, keeping in mind their intrinsic complexity, which reflects their essence, and should not be lost sight of, for fear of forgetting their constraints, operating rules and purposes;

– the second is that these sociotechnical systems by nature require a multidisciplinary approach. This approach finds its bases in engineering science as well as in human (ergonomics, psychology of behaviors, etc.) and social (anthropology, sociology, economy, etc.) sciences, to which it would be appropriate in many cases to add ecology (control of environmental impacts).

However, there is no one-size-fits-all system analysis formula, but a methodological approach that has to adapt to the types of systems under consideration and the circumstances or environments we will discuss. Systemism does not mean systematism! Although, in addition, systemic analysis makes it possible to build bridges between different disciplines and to provide them with fruitful insights, it does not replace these disciplines which naturally retain all their relevance, each in its own field. In other words, we note that the systemic approach offers a common analysis framework to all of these disciplines. It has also contributed to the development of new disciplines at the interface of existing disciplines. We have previously evoked control or information theories. This is also the case with biomechanics: this discipline developed at the interface of biology and mechanics, and has helped in making significant strides in understanding the behavior of the human body in the event of a violent shock, and has, in the field of transport, equally led to significant progress in passive vehicle safety.

Other classifications than that of Boulding have been proposed. They rely, for example, on the concepts of self-organized or regulated systems¹⁰. Their ambition is generally to seek the ways of some sort of unitary general theory of scientific knowledge, the very principle of which is the subject of debate. We do not intend here to take sides in this debate, but simply to examine the systemic method for its irreplaceable contributions to complex system engineering and, for our specific purpose, in the fields of town planning and transport. Therefore, we did not deem it necessary here to reflect further upon these different classifications.

Another advantage of Boulding’s typology is that it also allows us to make the modeling act more concrete. This will be discussed in section 8: it consists of striving to make the complexity of the object of study intelligible. To this end, a common but non-exclusive¹¹ process consists, in order to simplify n-level systems, of looking for analogies in (n-1) level systems or below. For example, social systems have been able to search for biological or ecological analogies (resort to Darwin’s theory of evolution). Yet, among the nine levels distinguished by Boulding, it is essentially the very first ones that have given rise to rigorous mathematical developments and models. There is therefore a strong temptation of having recourse to models derived from physics. Examples can be drawn from the transport domain: flow functions derived from fluid mechanics are commonly used to process flows, or gravity functions and more generally potential functions for quantifying the emission or the diffusion of these flows at different points in space. These analogies are not in themselves illegitimate¹², and should be judged on their capacity to report on observed phenomena. As we will highlight below, the modeler’s honesty consists of explaining in each case the model he/she uses, its underlying assumptions, and the consequences that may arise from it with regard to the validity of the proposed model. For example, the gravity model gives a good report on home-to-work commuting; it is, on the contrary, unsuitable for tourist travels, which proceed from other economic and behavioral logics. In general, the more detailed a model is, the more its scope of application is limited.

1.5. The concept of variety

Another way to assess the complexity of a system is to use the concept of variety, which was introduced by W.R. Ashby [ASH 58] from the bases of information theories¹³. For a simple system, it refers to the number of different elements in the system (the basic assumption being that any communication necessarily implies the existence of at least two possibilities). By extension, the systemic variety is measured by considering both the existence of the elements of a set and the combinatorial relationships between these elements (either for a number N of elements, a potential variety of N (N-1)/2, or N (N-1) depending on whether the relationships between the elements are oriented or not). If we are dealing with a dynamic system, whose elements and/or relationships between the elements can take different values or states, the variety is then measured by the varying number of states that the system can access. We realize that this number will increase exponentially with the number of elements¹⁴.

From our point of view, this concept is not aimed to provide us with a precise measure of the variety of sociotechnical systems. This claim would be obviously illusory and of little interest. Our main concern here, however, is to understand how the variety of a system evolves through certain actions, and how this concept can help us to understand the more or less serious difficulties to overcome in order to reach a sufficient control level (with regard, for example, to expected performances or a desired level of safety) of these sociotechnical systems.

It is useful here to resort to another interesting concept, that of constraint, also introduced by Ashby and conceived as a reduction of a system variety, thus the number of states to which the system can access. In an extreme case, where each element of each system is connected to each other such that a change in one part causes a change in all the others, the system variety is maximum. The introduction of constraints can come from the internal laws of the system (for example, physical laws such as the law of gravitation) or the result of an exterior action on the system. The organization thus created gives rise to a restriction of the variety. We have not developed this concept of organization – this will be discussed later using a few examples in section 1.7 – and it should be noted that it often refers to a static vision of the system. From a dynamic perspective, we talk instead of regulation, this being conceived as the action of maintaining a system as close as possible to a predefined state or evolutionary path. In the field of transport, regulation will consist (for example, regarding the railway) of taking any useful exploitation initiative to as much as possible comply with preprogrammed movement of convoys (and in particular their schedules).

For example, from these considerations, transport systems can be classified by order of increasing complexity. The most simple systems are those with maximum imposed constraints on the system: guided systems, subways, trams and trains that meet this criterion. Among these systems, subways, which usually have a protected infrastructure and function as carrousels, are the simplest. Tramways on their exclusive right-of-way come next: even protected, the infrastructure is in open air and its condition is therefore more difficult to control. This is even truer for trains, as the distances between the stations, that is, areas difficult to control, are much more significant.

At the other end, we find road transport: each vehicle has a significant autonomy, to which is added the difficulty to act on user behaviors with the vast majority of them not consisting of driving professionals. A driver’s inappropriate initiative may suffice to significantly change the state of the system (see, for example, the consequences of double parking). The use of rules or dissemination of information can induce forms of organization, but these are naturally less stringent than the physical limitation of the number of degrees of freedom of the system. The motorway system allows the reinforcement of constraints through entry/exit control and elimination of potential conflicts (separation of traffic directions, removal of intersections).

Between the two extremes of guided and road transport, there are other modes of transport (maritime, air, etc.) for which imposed restrictions and means of control correspond to intermediate situations. It can be observed that such classification of the complexity of the different transport modes reflects the statistical data on the rate of accidents that occur for each of these modes¹⁵: road transport is the most accident-causing, while rail transport presents the least risk exposure¹⁶, subject of course to a proper maintenance of the technical system¹⁷.

This example shows that complexity, from our definition, is not to be confused with the technical complexity of the system: the road system, which may technically seem the simplest, is also the one which, taken as a whole, is the most complex. This apparent paradox is a consequence of a basic rule in control theory. According to this rule, a system of variety V may only be totally controlled by another system if the variety of the latter is at least equal to V. In other words, in the field of transport, an activity whereby nature systems are largely open to their environment, control involves both technical sophistication and a significant reduction in the controlled system variety. In particular, automation, which is the most accomplished form of control, involves, at least with currently available technologies, a significant reduction in the initial system variety. Therefore, it will be easier to implement automation for the least complex systems, such as subways, whose relationships with the environment have already been simplified to the maximum¹⁸. It can even be said, in this approach, that the complexity of the transport system and technical sophistication vary inversely.

1.6. Keys to analyzing a system: functions and structures

We highlighted the interest of an inductive approach to understanding and/or modeling the functioning of a system. Systemic analysis offers us other equally valuable tools to progress in this direction.

It should be noted that although our definition of a system may not be precise, we can, however, specify some of the characteristic properties involved in this definition. Three of these properties seem essential:

– autonomy: beyond the observation that the boundary of a system with its environment is more or less vague (or defined by the modeler depending on the objectives pursued), a system is a set of elements which, because of the relationships binding them, has its own existence. Its behavior can therefore not be considered as (or reduced to) a direct (or linear) consequence of its exchanges with its environment, and this even though such exchanges are essential to the system (autonomy does not mean independence);

– coherence: this characteristic means that the system elements, which can be distributed in different points of space, behave in a unified manner: the evolution of each element can have consequences not only on the system as a whole but also on the behavior of other elements, which may end up being modified;

– permanence: it is the same characteristic as previously mentioned, applied to the time dimension. In other words, the system evolves (for example, to adapt to its environment) while preserving its autonomy and coherence. As individuals, we are autonomous systems that we will grow and age while maintaining what makes our personality and distinguishes us from each other.

1.6.1. The concept of function

Boulding’s typology allows the introduction of another important concept (still without presuming a demonstration here): the concept of function. Each level of this hierarchy of complexity can be read, beyond the first two levels that correspond to simple physical systems, as introducing one or more increasingly elaborate functions:

– level 3: the regulation and/or control function;

– level 4: reproduction function;

– level 5: adaptation to the environment function;

– level 6: teleological function (action responding to basic stimuli);

– level 7: intelligence function (consciousness and ability of adapting the environment to human needs);

– level 8: complex social functions: production, planning, political functions, etc.

It is intuitively perceived that there is a link between the properties of previously defined systems and the fact that these systems perform one or more functions: coherence, for example, stands as a condition for a system to perform a function, and conversely, this function can contribute to reinforce this coherence¹⁹ by constituting for the system a global objective capable of finalizing the behavior of each of its components. In this sense, some authors reserve the system terminology at level 4 and at levels above 4 in Boulding’s typology (according to Boulding, systems have a purpose, whether explicit or not). An obviously essential question – which could itself give rise to a system typology complementing that proposed by Boulding – is whether the function(s) associated with a system is or are imposed on it by its environment, or if it is or they are the result of a procedure internal to the system²⁰. This question, for the cases we are addressing here, is important: urban systems are constituted incrementally, by additions and successive transformations in connection with routines and practices, as much as they are the result of a design work.

This remark leads to another: as already observed for transport systems, we can, among sociotechnical systems, establish a certain hierarchy of complexity. Thus, industrial systems, which fulfill a unique if not main function, are less complex than urban systems, which are by nature multifunctional, and must reconcile partly contradictory functions: the same space, the street for example, must provide traffic, parking, delivery, access to businesses and riverside property functions, as well as more occasional functions (relocations, urban furniture repair, access to underground technical networks, etc.). With regard to the sewage network, it must evacuate wastewater, mixed with rainwater or not, while protecting the environment, etc. And we could cite many other examples.

1.6.2. The concept of structure

The function performed by a system depends on the nature of its elements and the relationships between these elements. The systemic approach will lead us to assume again that the link between properties and function(s) should mainly be sought in the analysis of relationships between the system elements, and thus in the system structure. A common method of system analysis consists of identifying this structure when it can be explained, or at least be the subject of empirically verifiable hypotheses, with the understanding that the relationships taken into account can be material (material and energy flows) or immaterial (information flow). Information flow analysis in general is of particular interest: this method is found in engineering or information theories as well as in certain branches of sociology (sociology of organizations in particular).

A simple illustration of this concept of structure in the field of territorial system engineering is in the analysis of networks that constitute infrastructures (transport, water and sanitation, energy, telecom, etc.) at different territorial scales²¹: here, the system is