University of Southampton OCS (beta), CAA 2012

Font Size: 
Modeling Anthropic Ecosystems: a framework to understanding the whole before modeling the parts
Andreas Angourakis

Last modified: 2011-12-16


It’s already a common place for archaeologists to interpret certain aspects of the archaeological record using the concept of ecosystem. Unfortunately, this practice became a convention as many authors ignore clear definitions, therefore leaving untouched significant implication of this concept for their interpretation. Maybe the most important implication of “thinking ecologically” is to conceive human beings as participants of a “natural” whole; natural as not being a by-product of individual intentional actions, at least not only.

Although it seems conflicting, the holistic attitude of the ecological thinking fits well with bottom-up approaches, for all ecosystems are formed and maintained by evolutionary processes. Thus, they are emergences of differential individual performances, constrained by abiotic elements and especially by the performances of other individual organisms, both co-specific and else, directly and indirectly, through dynamic ecological relationships. Actually, ecosystems are among the favorite specimens in the study of complex systems.

Given that all archaeological remains correspond to ecosystems that include humans, it should be more useful to archaeologists to refer to “anthropic ecosystems”, instead of using a broad notion of ecosystem. It is important to avoid the terms such as “man-modified ecosystem”, because it suggests a rough dichotomy between humans and everything else. However, one welcomes the use of other ecological concepts by some archaeologists (e.g. “niche construction”), because they allow us to frame some key roles of our species in whatever ecosystems we participate.

Since by definition all humans exists in an anthropic ecosystem, modeling human behavior –especially those related to subsistence— should be done within the “universe of possibilities” specified by the whole of a given anthropic ecosystem.

The study of a particular behavioral space is of course what makes computational simulation models so promising for the social sciences. Nevertheless, to escape from a plain accumulation of ad hoc models that can’t be directly compared with each other, a general framework could be proposed retaining the concept of anthropic ecosystem as a common prototype.

In this communication I will present a procedure to be followed before modeling multi-agent systems. This heuristic focuses on better understanding the anthropic ecosystem that we are interested by modeling unstructured populations dynamics with a dynamic systems approach. Is not the case to oppose DS and ABM modeling paradigms, expecting best fitting results from one of them for particular case studies. Instead, the purpose is to combine them directly, eventually designing a modular hybrid model, or simply by using the respective models as different instruments with different virtues and limits, that points to different potentially relevant aspects of the same question.

Even considering that evolutionary processes are strongly path-dependent, it is reasonable –and useful— to model ecological relationships of anthropic ecosystems first as deterministic (e.g. modified Verhulst differential equations) to further introduce discrete event and stochastic modeling techniques, towards designs more compatible with ABM. To exemplify this method, I will present preliminary models built on two cases: the cereal domestication in Near East and the interaction between farming and herding in oasis of Central Asia.