Computer-based Simulation Modelling for Anthropologists
Michael D. Fischer

Simulation Home










Simulation: Introduction

Simulations are useful for many purposes. You explore situations for which it would be difficult or impossible to do research for practical or ethical reasons, such as the impact on group sex composition of male-biased childbirth strategies, or the impact of famine on local political organisation. You evaluate the extent to which an anthropologist's account is consistent with the information they provide: do they provide sufficient information to justify their conclusions? You explore holistic interconnections between relatively simple principles when joined in a common context .

Simulations put models in context by letting us observe how the model behaves in a complex environment. 'Model' literally means to demonstrate or copy. In anthropology model more typically means some kind of representation used to account for some ethnographic facts. Anthropologists often use models to generalise the results of their research.

For example, the following is a very simple interactive simulation (e.g. a simulation where you can intervene, rather than simply sitting back and watching) of a situation that arises in Pakistan (and other Muslim societies) derived from Islamic Shariat law concerning a form of divorce invoked one spouse saying talaq to the other.

To run it click on the button marked'Say talaq' - the day counter shows the effect (or its absence) of the passage of time...

What's going on in this simulation? What are the conditions that result in divorce? What
social roles are indicated. What can we conclude from this?
(See the end of this page for a brief explanation)

Simulations of this sort are nothing new to anthropologists - we do them all the time in our
head or on paper to validate our analyses against data, to explore the properties of our
models, or to extrapolate our models to new situations.  Such simulation, predating
computers, has been used in anthropology at least since the nineteenth century  (Mulvaney
1970). Simulation, with caution and reservation, is used to observe rituals, ceremonies
and activities, which for some reason cannot be observed in the ordinary course of
fieldwork (Clammer 1984:72-3; Ellen 1984:274). Indeed, formal interviews and 'set-ups'
(Jackson 1987:41) meet this sense of simulation to some extent. Finally, we practice
simulation each time we 'play out' our models and analyses in our minds or on paper,
testing against observed data, and evaluating the results.

The use of computers for simulation modelling were among the first encounters of social
anthropologists with computing (Kundstater et al 1963; Gilbert and Hammel 1966), not
only because simulation met more or less the conception of what computers did in the early
1960s, but also because anthropologists at that time were beginning to explore the use of
more sophisticated models and attempting to apply a more systematic perspective to

What's going on in the Talaq simulation?

A rule adapted from Islamic Shariat law:

  If talaq is said three times in succession by a husband to his wife the marriage is
dissolved, otherwise the marriage continues.

and apply this rule to the following information - talaq was said twice by the husband to his
wife - we arrive at the result that the marriage continues.

If the wife says talaq three times to her husband, there is no prescribed result. If the husband
says talaq three times, the marriage is dissolved.

Although not represented here, there is much more to talaq divorce. Such a divorce is final. The
couple cannot remarry unless both are remarried beforehand.

Anthropology and Simulation

Anthropologists have been using simulation almost since the beginnings of the discipline. In the field, although it is most desirable to witness performances of ritual, preparation of materials etc. in person, this has not always been possible. More recently, since the introduction of computers, simulation has taken on another meaning; using computer models to explore social situations which could not otherwise be easily investigated. In its infancy simulation consisted of relatively simple numerical models which were iteratively executed towards a cumulative result, moderated by the influences of other models as the simulation run progresses.

Heretofore simulation has mostly been applied in a contextual way, though there are notable exceptions. By far the most important use of simulation in anthropology to date has been to evaluate the interrelationships between demographic structure and real or hypothetical social structures or cultural practices (cf. Hammel, Randolph and Coult, Dyke and McCleur). Simulation has also been useful in relating aspects of anthropology to policy (Nardi).

Buchler and Fischer have used simulation to choose between different models of land allocation for horticulture in New Guinea, Fischer and Selby used simulations to investigate the relationship between cultural models for agricultural planning and crop yields.

New techniques for modelling became available in the late 1970s in the form of knowledge representation by expert systems. Fischer and Fischer and Finkelstein have done detailed work on marriage arrangement in the Punjab, Read and Behrens on literacy in ...., Kippen modelled tabla improvisation, Benefer and .... the classification of land types, and more recently Fischer has been working on modelling the recreation of cultural traditions in the South Pacific.

These new approaches have not been properly exploited as yet. In part this has been due to the accessibility to the non-specialist of both knowledge-based techniques and the equipment needed to exploit knowledge-based techniques. There was also the necessary period of basic research required to evaluate the applicability of these methods to anthropological problems. Thirdly, there are newer prospects which promise to make knowledge-based simulation much more relevant to the discipline and its applications.

Simulation is the ideal platform to extend current knowledge-based methods to anthropology. As an actual model of cognitive processes KB models are not only unproven, but extremely unlikely. In terms of producing predictive classificatory behaviour that is comparable to those produced by actual cognitive processes, they are very successful. In particular Fischer's research demonstrates that a sufficiently detailed knowledge-based model produces classificatory results that are comparable to indigenous thinkers, and are acceptable to them.

Although these models are not as yet satisfactory from the point of view of representing the actual structure of indigenous thought, they are nevertheless useful in two capacities. Firstly, as a more formal representation method for ethnographic data. Secondly as components in computer simulations.

The typical structure of a simulation is a number of models of different order, which interrelate at the level of their behaviour. For example, we might wish to produce a simulation to investigate different methods of distributing active knowledge about Oral Rehydration Therapy (ORT) in a specific community. We will have available to us demographic data, traditional and formal educational facilities and activities, epidemiological data relating to relevant diseases over time, outlets for distribution, outlets for notices, medical ethnography which relates indigenous thinking about disease, clinical interview records, self-treatment strategies, sources consulted for medical purposes, etc. And, of course, ethnographic data which relates to these and other issues, in lesser and greater detail.

Some of these data can be represented using appropriate statistical procedures, linear programming models, and other numerical techniques. Others will be defined using algebraic or logical statements, represented in trees, graphs, lattices, or other systems. It would be an enormous task to a priori interrelate these data into a single formal system for analysis.

Simulation can be used as a tool for exploring just these connections. While the formal description terms of each kind of data may be of different order, we can develop a simulation model to investigate the interaction of these models to the extent the models can a) generate behaviours/results, b) other models can use these behaviours/results in generating their own behaviours/results, and c) we can state a useful problem in terms of these interactions.