1The preceding chapter sets out a brief overview of the different ways in which ‘transitions’ are approached in the scientific literature, from the point of view of both the natural and the social sciences. The theoretical aspects have been emphasized. The objective here is to render this concept operative for studying the changes in ‘settlement systems’ as it is possible to observe them, on the local scale of a microregion as well as on that of the planet. Ever since modern man has existed, he has used the surface of the earth, moved on it, exploited its resources and exchanged goods and information. He has thus exercised upon it a number of elementary functions, par lesquelles toute société transforme le milieu dans lequel elle vit en un espace organisé selon des normes humaines1 (‘by which every society transforms the milieu in which it lives into a space organized according to human norms’). The term ‘settlement system’ suggests that the interactions between places, on the one hand, and, on the other, between the individuals who live in them and their environments constitute a driving force in the evolution of these systems, and notably in the transformations of the related spatial patterns. Such systems, in constant evolution, are often transformed progressively, and their dynamics are such that they tend to reproduce the structure and the functioning of the settlement system. Sometimes the changes are more abrupt, and the dynamics may lead to the disappearance of certain interactions and feedback loops that make up a system. In such cases, a new system, with a mode of functioning and a structure radically different, may emerge. It is this radical change which is interpreted in this volume as a ‘transition’. The first concern of the programme TransMonDyn was to elaborate a generic conceptual framework that might be used for studying transitions occurring on different scales and situated within different spatio-temporal contexts. The second was to employ this framework so as to identify and describe transitions on the basis of empirical data. A third was to modelling these transitions in order to highlight their structures and find out any regularity in the changes observed. The transitions result, indeed, from complex processes which modelling makes it possible to explore. These different objectives result in producing dialogue between conceptual construction (what is a transition?), empirical observation (how does one identify a transition in the realm of observation?) and theoretical reflection (what processes underlie the transition?). The objective of this chapter is to present the collectively developed conceptual framework and succinctly to introduce the settlement systems and transitions which constitute the case-studies of this project. The detailed descriptions and the modelling of these transitions will be the focus of the chapters of Part 2 of this work (chapters 4 to 13).

The notion of settlement system: dynamics of reproduction and transformation

2Before addressing the issue of transition, it is appropriate to return to the notion of settlement system and the dynamics which are associated with it. Indeed, a change cannot be classified as a transition except to the degree that one observes a radical transformation of the structure and the functioning of the system under study.

3Settlement system is a notion that carries varying definitions according to the context, especially the disciplinary context, in which it is employed. A brief survey of the terminology makes it possible to illustrate these variants as the concept is applied in contemporary times or to an ancient era.

4As Jeffrey Parson reminds us,2 the concept of settlement system in archaeology (especially in North America) was defined in the 1960s.3 It is founded on studies of so-called ‘settlement pattern’: By settlement system we refer to the functional relationships among a contemporaneous group of sites within a single culture. Several units of data (seasonality, technology, features, subsistence pattern, the systemic index, triangle diagrams, and geographic data) will be employed, which, when combined with our inferences on the settlement pattern, should permit an interpretation of the settlement system of the Riverton Culture.⁴In this definition, while ‘system’ refers to the functional relations maintained by contemporaneous sites, the term ‘pattern’ applies rather to the way in which human beings make use of their space, how they occupy it. It refers to habitats, to their spatial arrangement, to their nature, as well as to the disposition of other structures necessary to communal living. 5In this sense, the distinction between ‘pattern’ and ‘system’ implies the opposition between a static perception of the organization of settlement (as embodied in the traces perceptible by archaeology) and a dynamic perception which takes account of social structure and the relations which have produced a given pattern.Parson refers clearly to a permanent habitat in inhabited nuclei. However, the distinction between ‘pattern’ and ‘system’ is also illuminating when it comes to analyzing a nomadic or semi-sedentary habitat. In that case, the term ‘settlement pattern’ refers to the spatial distribution of archaeological remains as observed in the present world. Those remains correspond to the habitats occupied, permanently or not, successively or at the same time. The term ‘settlement system’, by contrast, concerns the modes of functioning of a set of settlements considered together: relations between places, their complementary functions, the temporal extent of their occupation. It is in this sense that Lewis Binford uses the concept of settlement system,6 like Kwang-Chih Chang before him,7 distinguishing between the pattern of the spatial arrangement of the archaeological sites as they are recorded and the human strategies for occupying the landscape. By comparison with the American school, the English school proposes an approach to settlement systems that does not dissociate the spatial structure of habitat from the social organization linked with it and which implies relations between communities at once horizontal (cooperation, competition) and vertical (hierarchic, political).8 Moreover, this approach combines all the levels of study of habitat, from intra-site to inter-site, without precisely explaining what the entities are that intervene at these different levels, for everything is implicitly bound up together. Considering this difference between the American and English schools, the Canadian Bruce Trigger clarifies the debate by explicitly defining three levels of settlement:9 the individual unit of the habitat (the building), the community layout sometimes referred to as community settlement pattern (a community corresponding to a group occupying a space in association – multiple families), and the zonal patterns or regional settlement. At this last level are considered density and distribution of the population of a region dealing with their environment, the availability of resources that are being exploited and their relationships with other groups of inhabitants, human and non-human.When studies deal with the first societies possessing a state dimension, this integrated concept of a settlement system is adopted again, notably with the support of the models developed by certain geographers and economists, Christaller’s central place theory to characterize the pattern of settlement systems within space, and the rank-size10 law to characterize their hierarchic organization.11 From this point of view, the definition of the concept tends to approach that of geographers working on contemporary societies: a settlement system (considered here as synonymous with habitat system) is the ‘ensemble des noyaux habités (hameaux, villages, villes) et des réseaux de communication, matériels et d’information, qui connectent ces nœuds’12 (‘ensemble of inhabited nuclei (hamlets, villages, towns) and of the networks of communication, materials and information that connect these points’).In this same perspective, but in order to work on more than two millenaries, the authors of the volume Des oppida aux métropoles13 (a collective of archaeologists and geographers) define the settlement system as the ‘ensemble des lieux occupés d’un territoire ayant entre eux des relations évolutives d’interdépendance’ (‘entire group of inhabited places of a territory having evolutionary relations of interdependence with each other’). This definition makes use of the relatively neutral term ‘places’, which presents the advantage of being compatible with various periods and is independent of the type of data observed (statistical measures or archaeological vestiges). In addition, it explicitly foregrounds the evolutionary dimension of the relations between the entities under consideration.In TransMonDyn, we situate ourselves in the line of this work, while considering more diverse time scales. We have chosen to define the settlement system as a structured set of spatial entities various in nature (sites, catchment areas, farms, inhabited places, agglomerations, etc.), localized within a territory and having relations of interdependence (food subsistence, economic, social, political). This definition is more inclusive than the preceding one: it makes sense as a means of describing at the same time an ensemble of places, temporary or seasonal camps occupied by hunter-gatherers or a group of towns engaged in material and immaterial interactions. It includes dimensions of settlement that are both structural (spatial configuration, size distribution) and functional (relations between entities). This definition does not imply any a priori hypothesis concerning the durability of the entities considered or the existence of hierarchical relations among them.

5The relations between the spatial entities that compose a given settlement system are taken to be the driving force of its dynamic. Their form, their direction, their intensity and the feedback loops they present characterize the system’s functioning, which will be defined as its ‘regime’.14 The dynamics involved affect the system at several levels: 1) that of elementary spatial entities (e.g., sites, inhabited places); 2) that of sub-groups of spatial entities which function together (e.g., the inhabited places of a single cultural community) and constitute ‘sub-systems’; 3) that of the system taken as a whole, characterized by the relations between elementary spatial entities and between sub-systems. In the course of time, changes operate at each of these levels. Sometimes these changes are inter-related and simultaneously transform all the levels. The crossing of a critical threshold in one area on a given scale, for instance, can entail, by a cascading effect, the crossing of other critical thresholds on other scales.15 In other cases, by contrast, changes can occur at one level without involving transformations at another level of the system.The spatial entities composing the system may transform themselves in different ways – by changing in size (e.g., the number of inhabitants), in status (a city may become a regional centre) or in location (e.g., for a group of hunter-gatherers). They may also disappear or appear. Such changes at the level of spatial entities, however, do not necessarily imply a change in the structure or the functioning of the system. In other words, the elementary entities of the system may record transformations as radical as appearances or disappearances, while the system taken as a whole functions according to the same regime. The same feedback loops are regulating the system and tend to reproduce it maintaining its structure and mode of functioning. A settlement pattern, for example, can remain dispersed or concentrated overall, with functional relations of the same nature between settlements, when some of these have disappeared and others have appeared. Likewise, the spatial organization and the rank-size distribution of a country’s or a region’s cities may remain constant as certain cities decline and lose their places in the urban hierarchy, while others gain in these respects.

6Under the effect of the amplification of internal fluctuations or of an exterior perturbation (cf. chapter 1), the transformation of the structure is sometimes such that the system disappears and gives rise to a new system, characterized by a mode of functioning different from the preceding one. Such is the case, for instance, when cities having more complex economic functions emerge within a settlement system composed of hamlets and farming villages with relatively loose interactions. The interactions between spatial entities of a settlement system will then be of a different nature within the system of closely inter-connected cities which will emerge.16 These interactions will function in a way that defines a new ‘regime’. Another example would be the passage from a centralized organization to a polycentric one. Still another is the passage from a settlement system of hunter-gatherers pursuing a residential mobility strategy, consisting in changing their base-camp when resources are exhausted, to a system of hunter-gatherers who employ a logistical mobility strategy consisting in increasing the radius of the resource exploitation area, making use of secondary sites or practising storage techniques to improve access to food supplies.17 18In each of these cases, it is the very nature of the interactions between the spatial entities that changes, and thus it is the ‘driving force’ of the system, determining its ‘regime’, that changes. The objective of the project TransMonDyn is to study transformations of this kind. The examples just introduced concern scales and registers which seem a priori far from comparable. Nevertheless, in each case, there is a ‘before’ and an ‘after’ radically different with respect to the societies’ ways of occupying space. The unity of the project TransMonDyn does not reside in the choice of a given geographic region or period but rather in adopting an approach in which the settlement system is central and in privileging the spatial dimension of that system. The definition adopted above makes sense, indeed, as much on the local as on a world-wide scale, in the era of the hunter-gatherers as much as in contemporary times.

Construction of a conceptual framework for identifying transitions from the thematic point of view

7To identify a change in terms of transition requires a precise definition of the concept and of the methods of observation and analysis in order to deal with empirical data. Indeed, a transition is not spontaneously observable in the empirical world. A transition is ‘constructed’ by researchers and accordingly corresponds to an interpretation based on what they observe, on the one hand, and, on the other, on the conceptual framework they have adopted. Our objective has therefore been to design a common conceptual framework, which would be meaningful for transitions of different natures and origins in varied places and periods. This generic quality has proved indispensable in order to render the transitions comparable between themselves, independently of their geographic and historic contexts. The conceptual framework has been collectively developed and adjusted on the basis of the respective viewpoints and expertise of modellers and thematicians (linguists, archaeologists, historians, geographers). Two complementary steps have been identified:

• The first concerns the phenomenon of ‘transition’ as such: it involves identifying a change in the settlement system whose characteristics are such that the thematician who is an expert in the field proposes to interpret it as a transition. The classic demographic transition is a case typical of this approach.

• The second consists in identifying ‘qualitatively different’ regimes separated in time and interpreting as a transition the process leading from one regime to the other. It is, then, a question of characterising a first regime (by identifying a system and its mode of functioning), then a second regime (corresponding to a different system functioning in a different way). In so doing, the point is to locate and specify the dynamics that lead each of the systems so identified to reproduce itself. Moreover, this approach aims to identify the processes that produce the transition, that is, the passage from one system to the other. Some examples of such an approach are expounded in chapter 1, where it is notably shown how, for a given parameter, the crossing of a critical threshold can result in a change of regime.

8The first step served to identify the changes which the specialists of the geographical space and of the period interpret as transitions. The second step served to formalize each transition so identified according to a common conceptual framework. This framework is based on a systemic approach, on the one hand, and, on the other hand, on an ontological approach. Faced with a set of observations bearing on the characteristics of the settlement in which he or she is interested, each thematician has thus identified the entities and relations which formed a system in the sense previously discussed. The interactions between the entities composing the system represent the driving force of its dynamic. Over the course of its existence, as figure 2 of chapter 1 illustrates in highly synthetic fashion, the system evolves, fluctuates, even while maintaining itself around a constant general tendency. It is the feedback loops between the entities composing the system that regulate it and underlie its self-reproduction. If these, for a given reason (endogenous or exogenous) disappear, intensify or reverse themselves, the system as it has functioned ceases to exist. Entities of another nature, involving other types of interactions, may appear, and finally these new elements will be able to make up a system in their turn. François Durand-Dastès uses the term ‘systémogenèse’ (‘systemogenesis’) to designate this process in which a combination of elements present at a given moment within a given space makes up a system.19 Figure 1 illustrates some typical cases. At an early stage, certain of these elements may already be present but without being interconnected, or only weakly (the square with the question mark). Afterwards, interactions are created or become stronger, leading to the auto-reproduction of a system (indicated as ‘system 1’ in the figure). The disintegration of the system, following the disappearance of certain elements or relations, and the development of new relations between the elements present take place during a period of systemogenesis (marked ‘SG’) and result in the establishment of one or several new systems, which in turn evolve and auto-reproduce in their structure and mode of functioning. For François Durand-Dastès, the temporal ranges of the system and of the systemogenesis are different (figure 1): ‘Cette systémogenèse, de par sa brièveté relative, peut être considérée comme un évènement. On peut ajouter que cet "événement" est également un "avènement" dans la mesure où il introduit une nouveauté’.20 (‘This systemogensis, by virtue of its relative brevity, can be considered as an event. One may add that this “event” is also an “advent”, insofar as it introduces something new’).

Figure 1: Systemogenesis

9Novelty is a key to identifying the installation of a regime different from that which existed previously. In order to find elements of novelty and succeed in identifying a system ‘qualitatively different’ from another, we have relied on an ontological approach. For Barry Smith21, ontology is ‘the science of what is, of the kinds and structures of objects, properties, events, processes, and relations in every “area of reality””.22 Here, the ‘area of reality’ that interests us is the spatial pattern of settlement, so it is, therefore, a question of identifying the objects and properties that characterize it and the events, processes and relations that underlie its dynamic. An ontological change between two systems, then, can concern the appearance of objects, properties, relations or processes of a new kind. The diagram in figure 2 synthesizes the nature of potential changes between two systems. Suppose that each little circle represents a spatial entity of a settlement system: an initial system, corresponding to a regime marked R1, is represented with four entities connected, two and two, with one isolated entity present at the side of the system without interacting with it. Another system, corresponding to a regime marked R2, includes a new entity of a new kind (characterized by a new property), while interconnections, more numerous and varied, now link the group of entities as a whole within the space under consideration. It remains to characterize the processes involved in each of the two regimes, R1 and R2, to highlight differences significant enough to enable the passage from one regime to the other to be designated as a ‘transition’.

Figure 2: Passage from one regime R1 to another R2: changes in objects, properties, relations

10The conceptual framework developed in the project TransMonDyn relies on this double approach, systemic and ontological, and may be broken down into five steps (Figure 3):

• when observing their respective fields, researchers detect various ‘things’ (entities, relations…) and perceive evolutions. Figure 3a represents this step in highly stylized form: the lines indicate the trajectories of the ‘things’ so observed, some appearing in the course of time, others traversing the entire period under consideration, while still others disappear at a certain point;

• observation and analysis of data enable the researchers to bring out recurrent features and structures in these evolutions. These are indicated in very schematic fashion with red arrows (figure 3b) which represent hypotheses concerning the existence of relations;

• the facts observed are conceptualized by the researchers from the perspective of change, locating those entities that exercise a structuring function and identifying the interactions that ‘make up a system’ (black arrows). The cylinder of figure 3c schematizes the dialogue between the realm of observation, represented in the lower section, and that of conception, represented in the upper section.

• the researchers proceed in the same manner for a later period (figure 3d);

• the researchers compare the two systems, and if a qualitative difference exists, they interpret the passage from one to the other as a transition and investigate the processes which operate during this passage (figure 3e).

Figure 3: Schematic representation of the steps in identifying a transition in the settlement system

11The challenge of the project is to explore whether or not it is possible for thematicians coming from different disciplines and concerned with objects of differing types (groups of hunter-gatherers tens of thousands of years ago, Pueblo farmers, town-dwellers of Antiquity or today) to characterize the changes in settlement systems that interest them using this conceptual framework.

Identifying regimes and transitions in the empirical world

12In the empirical domain, when faced with a given issue, our objective has been to apply the different steps summarized in figure 3 in order to identify a transition in the settlement system and elucidate the systems corresponding to regimes R1 and R2. This requires assembling the necessary material resources (in terms of data, analyses, existing knowledge) to describe the structure of a settlement system and its mode of functioning in different time periods, then deducing its transformations from these. Beforehand, it is necessary to identify the entities making up a system and characterize their spatial pattern. Regarding contemporary periods, for example, if one wishes to study the evolution of the population of a city and of the dynamic of the system of cities to which it belongs, the entity ‘city’ must be defined with precision. There are several ways of doing this,23 however, and identifying a possible change will depend on the choice that has been made. For a settlement system in the past or future, quantifiable data rarely exist, and the challenge in that case is to construct or reconstruct the entities, estimating their importance (in terms of extent, population and activities), their localization, their relation to their environment and to other entities. In exceptional cases where vestiges have been preserved,24 the environmental and archaeological, even the textual, data may be abundant, and then it is a question of sorting and filtering these to obtain the necessary information from them.25 26In most of the areas inhabited today, however, the quality and quantity of the archaeological data may vary, in particular due to the field work possibilities27 28. In addition, except in a humid context,29 the paleo-environments are difficult, indeed impossible, to reconstruct. What is more, depending on the contexts and periods, there is little or no written evidence. Finally, for the most ancient periods of pre-history, there are only a few scattered indications, as a result of taphonomic processes30 or the rise in sea-level, which has, for instance, caused the disappearance of any traces that might have been left by groups of hunter-gatherers at the time of their departure from Africa 70,000 years ago. Analysis of genetic data makes it possible to compensate partially for that absence. Access to data and the nature of those data are therefore highly variable from one geographic area to another, from one period to another.As a result, it has been necessary to develop a research strategy which would enable the adaptation of a single conceptual framework (defined in the preceding section) to diverse settlement systems of the past, present and future, whereas the data are variable with respect to both their source and their nature. That is why, in the end, the characterization of settlement systems in the project TransMonDyn has relied on the expertise of researchers long familiar with the case studies chosen rather than on the elaboration of comparable databases. The observation and analysis of the data based on that expertise were effected prior to the project and the case studies of the project TransMonDyn were chosen on purpose. During the first phase, the specialists decided on the settlement systems their contributions would focus on. The changes which seemed interesting to interpret using the concept of transition as defined above were then identified. The second phase entailed taking stock of their existing knowledge in the form of organized and logically argued narratives. To ensure that these narratives might be comparable on certain points, they were uniformly organized according to the same protocol schematized in figure 4: the regimes R1 and R2 are identified using the five dimensions that characterize the settlement system. As is set out in the first section of chapter 1, these five dimensions correspond to universal functionalities which make it possible to describe every spatial pattern produced by societies across the world, whatever their political, economic and cultural forms, in the present as in the past. The thematicians thus described precisely the structure and functioning of each settlement system studied with respect to each of these dimensions in R1 and R2, focusing their descriptions on the elements most sharply differentiating the two regimes. For the operation of this procedure, the reader is referred to the chapters of Part 2. The next step consisted in determining collectively whether the common analytic framework made sense for a varied series of settlement systems and allowed for the interdisciplinary dialogue indispensable for effective comparative study.

13During the construction of the transitions, the thematicians have had to foster dialogue between the concepts brought to bear to describe a transition in the settlement system with which they are concerned and the knowledge, both empirical and theoretical, available to them regarding that settlement. Several questions were then raised: how to determine in practice that two regimes are different? Or, to put it otherwise, how to identify the appearance of new entities, properties or relations modifying the functioning of the system, in such a way that the change may be interpreted as the establishment of a new regime? How does one observe and measure these differences? These questions relate to the way in which time and space are apprehended in the conceptualization of a settlement system and to the kind of observations (indicators, measures) which can be applied.

Figure 4: Headings for a harmonized presentation of all transitions

Chronology versus temporality of the regimes

14In order to define regimes R1 and R2, it is first necessary to situate their existence within a time frame. The temporality of the functioning of a system can be described in two ways: by characterizing this functioning over a period (case 1) or at a given moment in time (case 2).

15The first case applies to several possible strategies:

• the thematicians focus their interest on a given historical period and examine the elements composing the settlement system during that period. In that case, the temporal limits are supplied a priori by a given chronological frame recognized as such;

• the thematicians may, on the contrary, prefer to ignore the classical historical divisions in order to take a ‘neutral’ view of the settlement system and allow temporal parameters to emerge for each of the regimes from the systemic analysis proposed above.

16The second case is useful when the thematicians are able to identify different modes of functioning at distinct moments, even if they do not possess enough information to propose dates for the beginning and/or the end of the functioning of the system corresponding to each of the regimes. Without having explicitly to define such chronological intervals, they can use the conceptual framework set out above to identify the regimes R1 and R2, in a relative succession of chronological frameworks, and interpret the difference between the two as a transition. The following question arises, in particular when the change is progressive: what accumulation of ‘little’ changes initiates a transition? The thematicians will often be more at ease when differentiating two situations that are clearly different than when precisely dating a rupture, a ‘mutation’. Indeed, to fix a mutation chronologically is all the more difficult because the five dimensions represented in figure 4, even if they are linked, may have variable temporal rhythms and experience shifts in their evolutions. In that case, a simple identification of some key dates, corresponding in a sense to the heart of the period of the system’s functioning, but without setting limits to that period, makes it possible to designate a transition (cf. chapter 11).

17Dating a change is a priori simpler when it has been caused in response to a particular and well known event or external shock (cf. inset 1). In certain cases, however, processes of change are already underway before that shock occurs and are not caused but simply accelerated by it (such cases are set out in the chapters of Part 2). Alternatively, a brutal event can trigger the transition, but the changes will actually take place several years, even several centuries, later. In all cases, it is necessary to specify the temporality of the change at issue, with its characterization in terms such as ‘abrupt’, ‘slow’ or ‘progressive’ depending on the time-scale selected.

18Finally, we should note that a similar type of transition process can operate at different chronological periods according to more or less variable temporality in different areas (cf. inset 2). Within the framework of the human and social sciences, an evolutionary perspective raises the question of growing complexity of the system studied over time. Such is the case, for instance, with sedentarization and the emergence of towns and cities, which have issued in an increased complexity of societies, with a growing number of entities, whose levels and especially interactions are more and more imbricated and diversified. Nevertheless, it seems to us important not to postulate such a growing complexity a priori. Accordingly, we clearly distinguish chronology, which presumes time as a succession of facts, and temporality, which refers to a rhythm of change independent of the position of the system along the axis of time. Thus, the decision to focus on the temporality of the process rather than its chronological definition enables comparison of case studies similar with respect to the driving force of change, regardless, moreover, of the socio-environmental context in which that change is produced. Consequently, the transition may be approached in the form of a ‘stylized fact’,31 something which facilitates comparison between cases situated in different spatio-temporal contexts (cf. Part 3). While adopting this methodological approach, we have preserved at every step of the work, from identification of the transition to its modelling, a characterization specific to each case study, so as not to give way to a simplification disconnected from any context. Indeed, even if the same mechanism may be replicated in different places and at different time periods, the factors responsible will only be significant in relation to the context in which the change has been produced. Spatial patterns and regimes

19The distinction between two regimes of a settlement system is founded on observable differences in the spatial configuration and interactions existing between the entities of the system. To illustrate our discussion, we will be drawing on an ensemble of theoretical cases which take into account the diversity of the situations encountered in the sphere of observation (figure 5).

20When observing the spatial configuration of a settlement system, the appearance or disappearance of spatial entities within the system represents a first indicator of change. Nevertheless, these appearances and disappearances do not systematically imply a significant modification of the system’s functioning. So, for instance, in figure 5a, even if some of the spatial entities have disappeared and new ones are manifest, the spatial configuration and the interactions which underlie the functioning of the system are not necessarily modified. In that case, we will consider that we are still dealing with the same regime, and that no transition has occurred. One may also observe a modification of the spatial entities, with variation in the size of the settlements, yet without any change in the functioning of the system. By contrast, observation of a change in the way in which space is occupied (5b) is a strong indicator of the existence of a transition. This is for example the case with nomadic groups, represented in figure 5b as ‘trajectory’ entities, which settle in a place, leading to the appearance of a new kind of entity: a sedentary settlement. The same is true when the appearance-disappearance of sedentary settlements presents a noticeable pattern characterized by a global shift of the point pattern (5c). Moreover, an appearance-disappearance kind of process does not only affect the spatial entities but may also be related to the way in which those entities maintain – or not – relationships with each other (5d). Thus, whether one is dealing with the trajectories of nomadic groups or fixed places of living, two distinct regimes may be considered as soon as autonomous entities develop interactions rendering them interdependent, whereas they more or less ignored each other before. Such conclusion does not pre-judge the intensity or the meaning of the relations that unite them. These interactions may subsequently evolve, moreover, gradually or rapidly, within the second regime. Finally, the passage from one regime to another may be expressed by a concomitant change in the relations and properties of the entities, one which underpins a significant change in function within the settlement system. Such is the case, for instance, with a passage from a practice of agro-pastoral subsistence to a ‘speculative’ one, consisting in generating surplus and stocking it for redistribution. In the latter instance, even if relationships already exist between the communities, their nature changes by virtue of a modification of the modalities of exchange, which tend to intensify and especially to become more unbalanced (5e). Such qualitative changes may illustrate a change of regime.

21The phenomenon of urbanization is part of this latter type of process, but its temporality may vary according to geographical areas and time periods. The city, or the urban character of an entity, supposes that there is a surplus of accumulated resources sufficient to maintain, on a collective basis, specialized persons within the community. The quantity of surplus a society is capable of storing is therefore an indicator of its ‘urban’ character, or of its degree of hierarchic organization. That quantity may evolve in a progressive manner or by sudden leaps, but the crossing of a critical threshold (cf. chapter 1) is necessary for the storage to become sufficient to free up the means to maintain persons specialising in functions other than subsistence production. The thematician may have to choose, in that case, between conceptualizing one or two transitions:

• In the first regime enough surplus would allow exchanges between weakly interconnected entities (with simple reciprocal relationships, for example);

• the second regime would appear after the crossing of a first critical threshold, whereby the level of storage becomes sufficient to practice economic exchanges. This new mode of exchange leads in general to inequalities in the redistribution of resources among entities and to the emergence of certain individuals in charge of the organization of the community (control of the stock, defence, religion, for example). With rare exceptions, the community as a whole remains diversified and still participates in the work of subsistence production;

• a third regime, finally, might correspond to the crossing of a second critical threshold whereby the quantity of stocks is henceforth sufficient to permit the emergence of a population specialized in different complementary functions (defence, crafts and commerce). In this case, specialized individuals no longer have any role in providing subsistence. In practice, no precise data allows the quantification of these critical thresholds. Archaeologists must bring to bear a collection of indicators in order to determine what state a society is in, and the spatial pattern of the settlement constitutes a key one from this point of view.

22These few theoretical cases show that a transition in the settlement system cannot be characterized exclusively by changes in spatial configuration. It is by taking change into account according to the five spatial and thematic dimensions that the existence, or not, of a transition can be identified and specified.

Figure 5: Examples of transitions in theoretical spatial patterns

How can one identify transition in the observable world?

23According to the time periods considered and the areas under investigation, different kinds of indicators make it possible to identify the potential existence of a transition but their precision is highly variable. When data exist or can be estimated,32 33demography is an appropriate indicator to detect a transition, as an abrupt increase or decrease in the rate of growth may reveal significant changes in the functioning of the settlement system, whether caused by or resulting from it. For instance, the Neolithic Demographic Transition,34 marked by an increase in the birth rate, reflects progress in agriculture and the associated changes in land use and food supplies in particular. In the absence of such measures, it is possible to fall back on other quantitative indicators, such as the number of settlements occupied over periods of 50 or 100 years, for example. If the variations in such values do not necessarily translate into demographic terms, they do take account, notwithstanding, of significant changes in land use, with phases of concentration or dispersion of the settlement. This is often reflected by changes in socio-economic, even political, practices. For instance, for the period between 800 BC (Iron Age) and 700 AD (late Antiquity), the curves in the number of settlements occupied by century (cf. figure 1, chapter 9) have made it possible to identify cycles, while the ruptures in these curves (sharp increases or decreases) have been interpreted as potential ‘candidates’ for a transition.35Dealing with change consists also in a qualitative evaluation of the distribution of the settlements, for example by using a rank-size law to analyze settlements according to the size of their population.36 37According to the profile of the curve obtained, it is possible to compare different types of distribution (log-normal, convex or concave), which give a signature of varied economic and territorial systems, more or less integrated.38 3940However, such interpretations imply precautions for the ancient periods.41 Other indicators more directly concern the spatial patterns that make it possible to point to a change in the settlement system. To analyze them point pattern analyses are used. This may involve quite ordinary indicators, which do not take account of properties other than spatial ones, such as the distance to the nearest neighbour, density or distribution (indicators of concentration or dispersion). Other classic spatial indicators integrate in their calculation politico-administrative function, such as central town status, as a property of one or more settlements located in the vicinity.42 43Some rely on a spatial model connected to the central place theory,44 or to a gravity model. That has for example been the case for studying the urban transition (chapter 7). Finally, in other cases, a potential transition may be identified only on the basis of a historical model or a particular ‘event-based’ theory. It is assumed that this event a priori has an impact on the structure of settlement but this impact cannot be measured, or even observed, either because there are no, or too few, archaeological vestiges or because the data do not yet exist, in the case of a future transition (chapter 13). For example, a change in the political system, characterized by a decentralization of power, would potentially have an impact on the spatial distribution of the settlement system, leading, especially, to a dispersion of the settlements and a more regular pattern (chapter 10). As another example, an ‘event’ as important as ‘Out of Africa’ migration of modern humans 4570 000 years ago implies a modification of the system of settlement by the colonization of new areas (chapter 4).Different methods based on indicators, measures and/or theories may thus be applied for identifying a change which will potentially be interpreted as a transition. The challenge is to make it evident that one actually passes from one regime to another, according to the whole set of dimensions presented above. Although it is preferable, for a comparative approach, to identify a change on the basis of a common indicator, the reality of the territories observed is often too heterogeneous to make this possible. Nevertheless, whatever the method chosen, the subsequent collective descriptive effort to apply the five dimensions allows one to verify whether the change hypothesized, or observed by measurements, really corresponds to a transition. This approach helps also to be precise about the nature of the transition and to identify mechanisms which can be compared from one case study to the next. In TransMonDyn, rather than focusing on the chronological and spatial exhaustiveness and/or representativeness of the case studies, we have chosen to favour their diversity, as well as disciplinary expertise concerning the period and the area under consideration.

Construction of a corpus of twelve case studies of settlement system transitions

24The framework presented above has enabled identification and description of twelve case studies of settlement system transitions corresponding to varied areas, periods and geographical and temporal scales.

25The 12 cases figure in table 1. The titles name the principal process associated with the transition at issue, by way of the term describing its emergence or, more generally, terms referring to a spatial process operating on settlement (expansion, concentration, dispersion, agglomeration, coastal development). These twelve empirical cases can be categorized in different ways. Three groups are distinguished (figure 6).

Figure 6: Presentation of the case studies according to their chronology and their geographic scale

26A first group includes four case studies bearing on a transformation of the settlement system on a global scale: the ’Out of Africa’ migration of our ancestors, Homo sapiens, beginning about 70,000 years ago, which entailed progressive colonization of the whole of the planet; the concentration of populations and human activities on coastal strips at the expense of the rest of the land area (the process is running for more than a thousand years); the emergence of cities, which took place independently in four or five areas in the world at different time periods, about 3,000 years after the beginning of the Neolithic; the emergence (– underway at present) of polycentric metropolises, ‘mega-cities’, which correspond to a change of scale in the functioning of metropolitan centres, with networks of cities and towns transforming themselves into ‘mega-cities’.

27In each of these cases, the distribution of the population on the surface of the earth has altered. In the first two, this change took the form of a displacement in the spatial extent of settlement, passing respectively from Africa to the whole of the planet and from the interior lands towards the coast. This transformation thus affected the pattern of ‘full’ and ‘empty’ areas of the surface of the earth (figure 5c). In the two subsequent cases, the change involved the emergence of new entities, cities and ‘mega-cities’, which progressively diffused over the whole of the planet. Settlement systems composed of dispersed spatial entities of smaller size were thus transformed into more hierarchized systems structured by larger urban entities exercising greater attraction (figure 5e), the largest of them finally evolving into polycentric integrated urban regions. Each of these great changes of structure, which imply the appearance of entities of a new kind, entailed change in the functioning of the settlement system. The last transition, for instance, that which is in progress and concerns the emergence of polycentric metropolises, ‘mega-cities’, corresponds to a change in scale in the functioning of metropolitan areas. With the intensification of interactions among places, the networks which one can designate as inter-urban in the first regime are transformed into new entities, whereby the inter-urban interactions become intra-urban in the second regime.

Figure 7: Localization of the case studies and representation of their temporal scale

28A second group is composed of four case studies on the regional level: the expansion in Africa of Bantu hunter-gatherer-farmers towards the south beginning in 3000 BP (farming taking on more and more weight, in relation to hunting and gathering, in the course of time); the passage in north-western Europe from a settlement system dominated by a dispersed settlement of short duration (the typical situation in 800) to a settlement system of greater density and subject to the attraction of more numerous and larger cities (the typical situation around 1100); the emergence of embedded urban networks and integrated systems of cities in the context of the co-evolution of road networks and networks of cities in France in the eighteenth and nineteenth centuries; the implantation of settlement in the totality of the south-African territory in different phases (migration of the populations of herdsmen-nomads, European colonization, which led to emergence of the first urban forms, and finally the shifting of settlement’s centre of gravity from the coast to the interior of the country, with Johannesburg as its economic capital).

29In these four cases, the entities of the settlement system are either groups (hunter-gatherers) which change location (type 5b) or spatial entities such as settlements (agglomerations, towns, cities) whose size varies over time (type 5a). In the first and last cases, the spatial distribution of settlement has changed, with a rearrangement of full and empty areas (typical case c of figure 5) following population movements corresponding to different colonization fronts. The interactions driving the system are linked to the exploitation of resources, foodstuffs in the case of the Bantu expansion, minerals in the case of South Africa. In the two other cases, the driving interactions take place, respectively, between villages and centres of seigneurial or ecclesiastic power and between urban units. From a structural point of view, the settlement system becomes hierarchical, and from a functional point of view the interactions become stronger, with an effect of attraction exerted by the more important spatial entities (urban units) upon the smaller ones (figure 5e). Both these cases see the emergence of systems of cities linked by exchanges of information which increase in strength (with the growing rapidity of communication).

30Finally, the third group assembles four case studies focusing on a transformation of the settlement system identified on the micro-regional scale: the formation of villages and the emergence of ‘polities’ (political entities) in the Pueblo societies of the south-western United States (600/900); the concentration of habitat in the Iron Age (600 BC/400 AD) in southern France; the integration of territories in the Roman imperial system in Mediterranean Gaul (200 BC/200 AD); the fragmentation of territories and the restructuration of the settlement system in late Antiquity in the south of France (300/700).

31In these four cases, the settlement system is composed of inhabited places which interact through relations of a cultural, economic and political nature. Settlement occurs in a framework where the relation to the environment is essentially based on an agro-pastoral system, where the societies are sedentary and solidly fixed within the area studied. In each of the settlement systems more or less pronounced population movements may be observed which can affect the system’s structure, without resulting, however, in a radical abandonment of the area occupied.

32Whether with regard to the Pueblo villages between 600 and 900 or the proto-historic and antique settlement in Mediterranean Gaul between 600 BC and 200 AD (table 1), the dynamic of the settlement system is characterized by a process of concentration and hierarchization of the settlement (figure 5e). The result is a more or less regular territorial mesh whereby the territorial entities tend to merge more or less rapidly into wider entities.

33Even if the relation to the environment and the entities in question are similar, the process appears inverted in the case of the settlement system in late Antiquity, between 300 and 700, compared with the first three cases. Indeed, there is no phenomenon of concentration but, on the contrary, a process of fragmentation of power and dismemberment of the territorial entities in favour of a major restructuration of settlement around small local centres whose distribution is more even.

Table 1: The 12 settlement system transitions studied in TransMonDyn

34The information base including 12 case studies featuring ‘small’ and ‘large’ transitions does not correspond to a succession of transitions, and does not claim to be exhaustive. It is rather a question of studying regions for which excellent expertise exists and which, taken together, reflect extreme variety in terms of socio-cultural contexts and spatial and temporal scales. Because the main goal of the project was to test the general applicability of the concepts by confronting as many different contexts as possible, the diversity of the case studies was a deliberate choice.

Conclusion

35Each chapter in Part 2 of this book is devoted to a transition (table 1). Once identified and described according to the conceptual framework set out and discussed in this chapter, the transition has been modelled.46 While the conceptual framework is generic, the models are varied and they are subsequently compared in the chapters of Part 3. Certain of these models are devoted to the emergence of entities or of new structures, while others study the consequences for the settlement system of exogenous events (climatic, political or technical). In both cases, the point is to assemble information and hypotheses concerning the interactions which provide the driving force in the dynamic of the settlement system. Depending on the authors’ choices, these interactions are taken into account at the micro level of the inhabitants or groups (families, groups of hunter-gatherers) who live, move and exploit resources within the area studied, or, instead, at the level of the spatial entities themselves (catchment areas, production zones, domains, networks, hamlets, villages, agglomerations). Thus, the models developed are varied with respect to the nature of the entities considered and the hypotheses relating to the functioning of the driving force behind the dynamic of the settlement system. The fact that the same conceptual framework is used gives relevance to an a posteriori comparison of these models. Each coupling of a narrative describing a transition in the settlement system with a model allowing the exploration and simulation of its dynamic constitutes a rich source of comparison. It enables reflection concerning the data, the nature of the entities observed and modelized, and the implications of the different choices of formalization (cf. Part 2). At last, Part 3 offers different comparative perspectives (chapters 14 to 17) on the trajectories. They range from identifying a transition in the realm of observation to the conceptualization of the regimes of the settlement system that precede and follow it, to the exploration of the mechanisms underpinning it on the basis of its modelling.