Models in Archaeology

David L. Clarke

• Publisher: Taylor & Francis
• Year: 2014
• eBook ISBN: 9781317606178
• Edition: 1
• Topic: Ciencias sociales
• Subtopic: Arqueología

This major study reflects the increasing significance of careful model formation and testing in those academic subjects that are struggling from intuitive and aesthetic obscurantism toward a more disciplined and integrated approach to their fields of study. The twenty-six original contributions represent the carefully selected work of progressive archaeologists around the world, covering the use of models on archaeological material of all kinds and from all periods from Palaeolithic to Medieval. Their common theme is archaeological generalisation by means of explicit model building, testing, modification and reapplication. The contributors seek to show that it is the use of certain models in particular ways that defines archaeology as the practice of one discipline, with a set of general tenets that are as applicable in Peru as in Persia, Australia as Alaska, Sweden as Scotland, on material from the second millennium B.C. to the second millennium A.D. They assert that careful model formulation within archaeology and the cautious exchange and testing of models within and beyond the discipline provides the only route to the formation of the common, internationally valid body of theory which defines a vigorous and coherent discipline and distinguishes it from being a collection of merely regionally applicable special cases.

1

Models and paradigms in contemporary archaeology

D. L. CLARKE

University of Cambridge, England

Introduction

Models are undeniably fashionable and especially so in the primitive disciplines that range between the Arts and the Sciences – geology, geography, archaeology, anthropology, sociology, economics, psychology and aspects of biology. Indeed, in some quarters this vogue and this setting are, in themselves, taken as sufficient reason to dismiss the significance of the model-using approach. This repudiation only requires the additional evidence of the dangerous errors that have arisen from the mistaken use of bad models, an outline of their clear limitations and the assertion that scholars have in any event long known and tacitly used these procedures, and the case is complete.

The question is, however, not whether models are fashionable and dangerous toys of long standing, but why models are so fashionable in disciplines of this primitive kind, why are some models inadequate, and why do models generate such strong feelings if they are part of immemorial usage? If we have an old and dangerous component performing an important variety of tasks in a number of disciplinary machines, should we then ignore it because it is fashionable – or learn about it, improve upon it and explicitly develop its potential? Beach’s (1957, p. 9) remark that ‘the history of economics could be thought of to a very large extent as a history of misapplied models’ might be generalized for all academic disciplines but could also hardly be improved upon as a sufficient justification for a more explicit appreciation of models, their variety, their uses, abuses, capacities and limitations – fashionable or unfashionable (Harvey, 1969, p. 142).

What is a model? Models are pieces of machinery that relate observations to theoretical ideas, they may be used for many different purposes and they vary widely in the form of machinery they employ, the class of observations they focus upon and the manner in which they relate the observations to the theory or hypothesis. It is therefore more appropriate to describe models than to attempt a hopelessly broad or a pointlessly narrow definition for them. Models are often partial representations, which simplify the complex observations by the selective elimination of detail incidental to the purpose of the model. The model may thus isolate the essential factors and interrelationships which together largely account for the variability of interest in the observations; in this way the model may even share a similarity in formal structure with the observations.

If explanation in general, and explanation in archaeology in particular, is viewed merely as a form of redescription which allows predictions to be made, then models as predictive forms of redescription are essential parts of archaeological explanation. However, here we enter the domain of the Philosophy of Disciplines, more commonly miscast as the Philosophy of Science, and since this introductory chapter will make no extensive attempt to consider the essential background of patterns of explanation, the role of hypotheses, theories and laws, or the problems of causality and indeterminism, the reader is recommended to consult sources noted in references throughout this volume.

The relation between the model and the observations modelled may in general be said to be one of analogy, or in the case of logical and mathematical models more usually one of isomorphism (Clarke, 1968, pp. 59–60). The analogy implies similarity between the analogues in some respects and dissimilarity in others, since otherwise the analogy would amount to identity. The set of shared similar characteristics is conveniently called the positive analogy, the dissimilar characteristics the negative analogy, and characteristics about which it is not yet known whether they are similar or dissimilar, the neutral analogy (Hesse, 1963). In general, models serve as heuristic devices for manipulating observations and hypotheses; they may also act as visualizing devices, comparative devices, organizational devices, explanatory devices or devices for the construction and development of theory (Harvey, 1969, p. 141). Models are usually idealized representations of observations, they are structured, they are selective, they simplify, they specify a field of interest and they offer a partially accurate predictive framework. In this way a map will schematically present an idealized representation of a selected item and its distribution on a simplified projection of a map surface, or a classification system may provide a crudely predictive algorithm based upon the identification of selected key attributes, or a mathematical equation may symbolically express the interdependence of systems of selected variables from within an archaeological situation.

Why need the archaeologist concern himself with models? There are five main reasons, which may be briefly outlined:

(1) Whether we appreciate it or not our personal archaeological opinions, approach, aims and selection of projects are controlled by largely subconscious mind models which we accumulate through time. We should realize that we are thus controlled.

(2) Whether we appreciate it or not we always operate conceptual models in the interpretation of observations. We all resemble the Molière character who was delighted to find that all his life, unknowingly, he had been speaking prose. We should make these operational models explicit and testable.

(3) The construction, testing, verification or refutation and modification of explicit models is the essence of the empirical and scientific approaches – providing the progressive cycle by means of which fresh information and insight are gained and theory is accumulated. Observations, hypothesis, experiment, conclusions, fresh hypothesis, fresh observations …

(4) The existence of a model presupposes the existence of an underlying theory, since a model is but one simplified, formalized and skeletal expression of a theory – be it tacit or explicit – developed for a particular situation. A careful study of groups of models apparently expressing a common underlying theory for different situations may therefore help us to expose, and articulate latent theory in a palpable and widely powerful form (Harvey, 1969, pp. 146–7). Model definition is a route to the explicit theory which essentially defines a vigorous discipline.

(5) Finally:

Hypotheses are generated from the model expression of a theory.

Explanation comes from tested hypotheses.

Hypotheses are tested by using relevant analyses on meaningful categories of data.

Thus models are a vital element in all archaeological attempts at hypothesis, theory, explanation, experiment, and classification.

Model building is important in archaeology, therefore:

Because it is inevitably the procedure used.

Because it is economical, allowing us to pass on and exchange generalized information in a highly compressed form.

Because it leads to the discovery of fresh information and stimulates the development and articulation of general theory of more than parochial importance. (Haggett, 1965, p. 23)

However, we must be clear about the peculiarities and weaknesses of models. It is particularly important to realize that since models mirror only selected aspects of the observations then it is both possible, permissible and desirable to have more than one model of different aspects of a single situation – an economic model, perhaps, and a sociological model of the same set of assemblages, where a trivial observation under one model may become a central factor under another. Under this pluralist viewpoint there are many competing models for each archaeological situation, where none may be finally picked out as uniquely and comprehensively ‘true’. The interaction between archaeologists and their material thus reflects a continuous flowering of models whose criteria are both internally and externally adapted to the contemporary climate of thought, and often between which no rational choice is possible. Let a million models grow (Hesse, 1963).

Very well, we may have more than one model at a time, all ‘true’, and the devising of a new model does not necessarily mean that all others are wrong. However, this permissible pluralism only holds for models selecting different bases for analogy. Clearly there are good models and bad models, powerful models and restricted models. Bad models are most frequently misused models whose function has not been clearly specified and controlled, or where the model has been over-identified or unidentified with theory, or that display a lack of clarity about which are the positive, negative and neutral characteristics of the model analogy (Harvey, 1969, p. 160). Although some models simply present pluralist alternatives based on the varying selectivity of the analogy, others will offer successive approximations of greater power and in that sense may be ranked relative to one another for a given function upon a chosen field, and a choice may be made between rival models in such a way that successively more powerful models are employed. Even the exercise of this choice between models on the basis of relative power is complicated by the many dimensions of model power which combine comprehensive, predictive, efficiency and accuracy elements.

Comprehensiveness	=	the size of the set of situations to which the model is applicable.
Predictiveness	=	the number of bits of information that the model can predict about individual situations in that set.
Efficiency	=	the parsimonious capacity of the model to make the most predictions using the fewest statements and the most elegant structure.
Accuracy	=	the quality of goodness of fit of the model predictions to the observations.

Controlling models

Amongst the reasons offered as justification for an archaeological concern with models it was implied that the archaeologist is the victim of a set of controlling models which affect his behaviour, including the discriminatory selection of the operational models which he chooses to deploy against archaeological observations. The nature of these cognitive or controlling mind models is both complex and composite. Through exposure to life in general, to educational processes and to the changing contemporary systems of belief we acquire a general philosophy and an archaeological philosophy in particular – a partly conscious and partly subconscious system of beliefs, concepts, values and principles, both realistic and metaphysical. These beliefs are then more consciously related to certain aims or goals by the mediating effects of our archaeological philosophy and its values, upon a series of alternative paradigms and methodologies. Thus, Kuhn (1970), in his fundamental work on the structure of changes in intellectual disciplines, has described the paradigm class of supermodel as implicit in the behavioural conventions held by a group of practitioners which delineate, focus upon and recognize a limited subset of data and experimental achievements, within their much wider disciplinary field, which for a time emphasizes the significant problems and exemplary solutions for that community of scholars.

The overall model of an archaeologist at work, therefore, may be represented as a set of controlling models which are embodied in the archaeologist’s philosophy, the paradigms he chooses to align himself with, the methodologies that he finds most congenial and the aims that this system constrains. The controlling models encapsulate the archaeologist and his chosen operational models, although there is feedback between all these changing elements. In effect the archaeologist is operating in a plastic sack and from within it seeks to push selected operational models against the complex archaeological reality, constantly seeking for goodness of fit, but the constraints of the controlling models always remain to obscure his perception of the reality (Fig. 1.1).

Archaeologists, like most other practitioners, have the greatest difficulty in believing that their own perceptions are controlled to this degree – they may always apply to other archaeologists, of course. But the history of archaeology and science is full of examples. Archaeology has witnessed the successive difficulties engendered by controlling models that could not accommodate the acceptance of the human manufacture of stone tools, an antiquity of man before 4000 B.C., a Palaeolithic origin for cave paintings, or tool-using apes 2 000 000 years ago. Medieval man, with his theocentric and geocentric controlling model of the universe, similarly resisted the scientific and heliocentric model of Copernicus and Galileo even in the face of evidence that the old model did not fit the new observations. In the eighteenth century, observed eccentricities in the orbit of Mercury were accounted for, using Newton’s model of the universe, by the presence of an unknown planet, which was promptly named Vulcan and sighted and described several times. The eccentricities are now accounted for in terms of other phenomena and the planet Vulcan has never been seen since. Even the direct photography of Mars has failed to convince observers who have spent a lifeti...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• List of Figures
• List of Plates
• Acknowledgements
• 1 Models and paradigms in contemporary archaeology
• 2 The methodological debate in contemporary archaeology: a model
• 3 Contemporary model building: paradigms and the current state of Palaeolithic research
• 4 Early phases of human behaviour: models in Lower Palaeolithic archaeology
• 5 Research design models
• 6 A model for classification and typology
• 7 What mean these stones ? Ethno-taxonomic models and archaeological interpretations in the New Guinea Highlands
• 8 Introduction to imaginary models for archaeological scaling and clustering
• 9 Models, methods and techniques for seriation
• 10 Computer models as tools for archaeological hypothesis formation
• 11 Initial model formulation in terra incognita
• 12 Socio-economic and demographic models for the Neolithic and Bronze Ages of Europe
• 13 Ecosystem models and demographic hypotheses: predation and prehistory in North America
• 14 Energy and ecology: thermodynamic models in archaeology
• 15 Ethno-historic and ecological settings for economic and social models of an Iron Age society: Valldalen, Norway
• 16 Ethno-archaeological models and subsistence behaviour in Arnhem Land
• 17 A computer simulation model of Great Basin Shoshonean subsistence and settlement patterns
• 18 A territorial model for archaeology: a behavioural and geographical approach
• 19 Set theory models: an approach to taxonomic and locational relationships
• 20 Locational models and the site of Lubaantún: a Classic Maya centre
• 21 A provisional model of an Iron Age society and its settlement system
• 22 Locational models of Transvaal Iron Age settlements
• 23 Locational models and the study of Romano-British settlement
• 24 Settlement and land use in the prehistory and early history of southern England: a study based on locational models
• 25 Models in medieval studies
• 26 Scientific inquiry and models of socio-cultural data patterning: an epilogue
• Index