Modelling Past Rural Environment-Society Systems

8.2.1 Distinguishing Modelling Objectives, Subjects

and Objects

Following Lieurain (1998); Garneau and Delisle (2002) or Bousquet and le Page (2004), three goals are assigned to modelling strategies: describe, understand and predict.

This trend of modelling the past is highly promising because it will help to solve large and theoretical questions, once parameterization and calibration have been established robustly to avoid the questioning over assumed postulates. One of the most famous simulations of the past, “Understanding Artificial Anasazi” can be categorized into such a scheme at a local “terroir“^[7] level: from a study (Kohler and Carr 1996), several theorized modelling experiments were assessed (Dean et al.

Moreover, and following Landais and Deffontaines (1987), a model is a “theo­retical and finalized representation of a reality formulated on the basis of situated observations, of a predefined framework that will then be applied to study cases and permits to give representations quickly”. A model serves to establish structural relationships and functions existing between the factors that one would like to analyse, and their respective importance (Parker et al. 2003). All models are designed for reaching a goal, an objective that can be problem solving, decision-support or simply experimentation (Roy 1992; Le Bars 2003). This definition adopts an oper­ational standpoint, thereby insisting on the subjectivity and the need for an objective for every model. In our case, this should establish what research question is inves­tigated with a model, i.e. what research subject and what research object is considered:

• If the subject is the history of the impact of the human expansion over an area, which can be either the Earth itself, a portion of it or a “terroir”, the studied object is the area, including the related natural resources, from which humans are seen solely as a transforming force, whatever the refining of the behaviour of this force can be (inclusion of innovations, etc.). The main use of such models is in environmental disciplines.

• If the subject is the history of the population itself and the impact of the environment over its evolution (again regardless of its size), environment and related natural resources should be considered as an influencing force, even if humans themselves transform the capacity of this force, and the research object is the population itself. The main use of these models is in (pre-) historic disciplines.

8.2.2 Agent-Based Modelling Distributed Simulation: An Adequate Tool for Modelling Villagers and Fields

Humans form complex groups and societies that are bound to their environment in more or less intense interactions, the imprint of which are found in landscapes.

Thus, archaeological/palaeo-environmental models can either directly analyse the social interactions between agents, or use the landscape as a reference plane. The choice of the adequate modelling technique is strongly dependent on the research question, and in many cases also on the scientific background of the modeller. In any case, it is the mutual interdependence of humans and their envi­ronment that is in the focus: environment and natural resources are quickly and directly affected by human activities and at the same time, humans are directly and rapidly affected by the availability of natural resources. We see here the interest and the efficiency of agent-based modelling tools. In the Neolithic, power processes creating a public policy resulting in a large transformation of the access to and/or the nature of natural resources did not exist.” Instead, such interactions are ato­mistic, i.e. they correspond to the repetition of small and direct transformations and uses of a territory at a lower spatial scale because of one or a small group of humans. De facto, they correspond well to the distributed way of conceiving large transformations of a landscape by repetitions of actions played by a multitude agents and actors that multi-agent modelling are the best to deal with. However, the large debate between descriptive and theoretical models initiated by Edmonds and Moss (2005) does have impact on the way multi-agent modelling is used for studying archaeological/palaeo-environmental issues.

8.2.3 Scales and Discipline Drivers: Categorizing Which Model You Are Working on

While a modellers’ dream can be to construct global models that integrate data of all relevant research disciplines, more local reconstructions with a narrower focus are better suited to meet the needs of local to regional heritage management.

i. Scales:

(a) The level of the village/hamlet (defined here along the more adequate word “terroir”) unit is often used because it is the functional unit of management of a landscape, the geographic expression of a combination of rationalities that have to interact altogether. Building a model of one simulated entity below this level is impossible regarding the importance of such interac­tions, both direct (marriages & other social interactions but also mutual manpower support for instance). Roughly, it is the level in which micro-economic rationality can be considered in order to analyse and explain differences in the use of natural resources;

(b) The level of the territory that corresponds to a culture or a group of cul­tures. Roughly, it is the level in which macro-economic rationality can be assessed, assuming a certain homogeneity regarding the use of natural resources within this culture comparing to others. A main aspect here is to analyse the impacts of a homogenous use of these resources;

ii. Involved disciplines and drivers:

(a) Archaeology and social science is used as the major input for the model and the expected results are focussing on palaeo-environmental issues;

(b) Paleo-environmental data are used as the main inputs for the model and the expected results are focussing on archaeological issues.

8.2.4 Genericness Criteria: Analysing the Validity Extension Methodology

Finally, we introduce a classification of different genericness criteria that models tend to achieve. In contrast to the binary categories of scale- and discipline-we describe above, (a model is either built on a rationality on a global scale, then tested on a lower scale - or the opposite; but cannot be both; a model uses disciplines as inputs and cannot therefore use the same inputs in a validation step), these genericness criteria are gradients on which one can position its model:

1.

2. the social-temporal genericness (innovations, adaptations, social evolutions): The very same problem concerning space may be applied regarding time as well: one may have to introduce evolutions of techniques, practices and/or social relations that can adapt themselves along simple (reactive, elimination) or complex procedures (learning, cognitive adaptation, etc.) to get out from the “instantness” of these models;

3. The Micro adequacies: local emergences and social differentiations. Following Fraser (2003), Misselhorn (2005) or Olivier de Sardan et al. (2007), once a famine or any other plague occurs, they affect only portions of the population (families, groups), mainly the most fragile ones, and not the whole population. This means that only very specific and catastrophic “plagues” (for instance well-referenced and very harsh droughts), may constrain simulated populations because they affected the whole population (see Dean et al. 1999;Axtell et al. 2002; Janssen et al. 2003; Kohler (2008) on the Anasazi collapse). More generally, any social and economic dynamic may not be seen as affecting indifferently a whole population, but only portions of it, combining specific parameters (for instance at the level of a family, lack of manpower, low cropping surface per capita, bad gender repartition regarding inheritance access, etc.).

One can then position a model's genericness extent and build its validity extension along these genericness criteria.

8.3