Archeologia e Calcolatori 13 - 2002

One of the tasks of cognitive archaeology according to C. RENFREW (1994) is «to use the well-established techniques of rational scientific inquiry, and to aim to develop these [...] by explicit theoretical formulations». Such is the purport of the ongoing research program initiated in France in the '70s on the logicist analysis and computational modelling of archaeological constructs (GARDIN 1979). A first assessment was presented to UISPP Commission 4 in 1990; the present paper describes advances of the program after that date in two directions, theoretical and pratical. 1. On the theoretical side, (a) new light has been shed on the position of the logicist analysis of archaeological papers (irrespective of their subject or denomination) in relation to recent work on natural logic or natural reasoning in the sciences of man; (b) further, the modelling function of the proposed 'schematisations' of argument has been brought out in the course of an ongoing debate on the respective part of Models and Narratives in the constitution of knowledge in the social sciences. The constraints to which mathematical models are currently subject are applicable to logico-discursive models as well: the same tests (formal coherence and empirical correspondence) are used to establish the validity of both; (c) lastly, as a logical follow-up of a and b, the case for a 'séparation des genres' has been strengthened, i.e. scientific models on the one hand, whether quantitative (mathematical) or qualitative (logicist), and/or imaginative amplifications of their findings on the other, both genres being however regarded as contributions to knowledge in a broad sense (BRUNER 1986). A large part of our discursive constructs belong to an intermediate or hybrid kind which tends to claim a distinct epistemological status, between or above the two genres. Doubt are raised about the future of this perspective in the long run; they found some unexpected support in Paul Ricoeur’s recent plea for a return to a stricter distinction between the cognitive and the rhetorical components in our «writing of history and representations of the past» (RICOEUR 2000). 2. On the pratical side, a new form of archaeological publication has been proposed (ROUX 2000), combining the principles of logicist analysis and new information technology. It consists in reformulating linear discursive constructs as tree-like structures of inference, expressed in computational terms (data base + rewrite formulas), and recording them on an electronic support, CD-ROM or web site, in order to take advantage of the navigational facilities of hypertext. No loss of cognitive substance is incurred in the process; and a partial answer is thereby offered to the 'reading vs. consultation issue' now widely acknowledged in scientific research, in the humanities as elsewhere.

Paleolithic art is a fundamental tool for the understanding of prehistoric societies. The relationships between paleolithic art and archaeological sites have been investigated and show the existence of nearly all the types of artistic expressions (sculptures, paintings, engravings, drawings, clay modelling, mammoth bone assemblages, etc.) in various sites like rock shelters, open-air sites, burials, deep caves, open caves and rock open-air sites. Unfortunately, old discoveries and tourism have destroyed in most cases the relationship between mobiliar and cave art with archaeological structures, limiting to recent discoveries the capability of a global approach. The different interpretative theories of prehistoric art since XIX century have been remembered both for symbolic explanations (Reinach, Breuil, Bégouen, Raphaêl, Laming-Emperaire, Leroi-Gourhan, Sauvet, Vialou, Clottes) and for social explanations (Efimenko, Abramova, Semenov, Iakovleva, Sieveking, Conkey, Bahn), and their revision due to the recent 14C AMS dates directly obtained on paintings made by charcoal. The critical question of the building of a reliable chronological framework is discussed. A method to study prehistoric art is then proposed, in five main steps: Step 1: Acquisition (recording); Step 2: Acquisition (signs, species and scenes determination); Step 3: Structuring (craftsmen workflow: space selection, physical-chemical studies, stylistic analysis, panel organization); Step 4: Structuring (chronological and spatial organization of the decorated space, relationships between the decorated space and the territory of hunters-gatherers); Step 5: Modelling (the craft system, the social system and the symbolic system of the hunter-gatherer society).

This paper will focus on the origins of the city. This subject has been studied in sociology, anthropology, history and geography, but there is not a unified approach. Our paper deals with the specific way social theory can be used in archaeology. We consider that a 'city' is a specific form of social space 'produced' by a series of social actions. However, this 'production process' cannot be described easily in archaeological terms. As a result, there is a deep gap between social theory concepts and archaeologically observable evidence. Today it is fashionable to speak about the unscientific nature of Archaeology and Social Science. This paper deals with this discussion, trying to create an observational theory to understand the process of city formation. We reject traditional positivist approaches of concept and reference, because of its simplicity. However, this fact does not mean that the analysis is impossible in scientific terms. We show how to use spatial statistics, probabilistic modelling and visualization technology in order to obtain a simulation of the spatial process, and then use the resulting model to build a representation of social theory in archaeological terms. In the paper we use the Italian city of Tarquinia as a case study. It is suggested that the origin of the city can be represented as a spatial process beginning with preliminary scattered villages, which join together forming bigger spatial units, which become attractive for the better geographical and geomorphologic conditions. The gradual consolidation of the main settlement in the best location is determined by the population growth, and the development of a new productive system and new social relationships.

In the last fifty years, many types of models on ancient social evolution have been created, both in the Old and New Worlds. This paper reviews the most influential ones, trying to summarize the recent, radical changes in the theoretical perspective on the emergence, development and collapse of complexity in human societies. The most serious problem, today, seems to be an enormous gap between the inadequacy of the archaeological record and the growing refinement of theories.

The use of ethnographic analogy to interpret archaeological remains has produced many misunderstandings, which must now be corrected. 1. Ethnology is traditionally oriented towards the analysis of the thought systems of the populations under study, and believes all too often that this type of discourse in natural language is an acceptable explanation for the observed empirical phenomena. The scientific discourse built by the ethnoarchaeologist must not imitate the distinctions made by the people under study, given that their constructs seek to satisfy different objectives. 2. The construction of inference rules must be subjected to the requirements of all scientific research. In consequence, one must not merely collect 'cas d'espèce', but also assemble numeric data which are sufficiently representative for statistical treatment to be carried out. 3. It is necessary to define, in each case, the actualisation context of the rules, that is, the spatial and temporal universe wherein the proposed rule is applicable. 4. The only way to validate a rule resides in the precautions taken during collection, mobilisation and treatment of empirical data. 5. Successful application of an actualist rule to archaeological data does not mean that the latter has been validated. The only way to confirm an interpretation is by applying the principle of result convergence by independent methods.

An attempt to build a global cognitive theory in Archaeology is proposed. The archaeological method is based on a three level concept : knowledge acquiring, structuring and modelling, inspired by the XIX century work of Peirce, renewed by recent developments of cognitive Sciences and used today in many fields of Social and Human Sciences, System Engineering, and recently proposed in Archaeology (DJINDJIAN 1993). The knowledge acquiring level A is the result of simultaneous and retroactive use of two mechanisms: several specific analogical processes in archaeology (contemporary analogy, ethnographical analogy, experimental analogy) and a cognitive process, general to Human Sciences. Logical objects used by archaeological reasoning are artefacts, set of artefacts (archaeological layer, dwelling structure, burial, etc.) and methodological objects (unit, sample, core, etc.). Such objects may deliver three categories of data: intrinsic data, extrinsic data and administrative data. Intrinsic data (named I) are a view of an object, resulting in the interaction between the archaeological artefact and the archaeologist who is perceiving and describing it. Intrinsic data is a knowledge of the artefact. Extrinsic data (named E) are data recording the various artefact contexts: spatial and stratigraphic localisation, links with neighbouring artefacts, environment, etc. Extrinsic data depends on the quality of archaeological excavation and recording. In all the cognitive processes, knowledge A must be associated with the archaeologist, ARCH, who is at the origin of the interaction artefact/archaeologist, the process of producing the knowledge, Pc, and the validation process Pv, controlling the reasoning: (A, ARCH, Pc, Pv). The structuring level, S, is discussed in relation with the question of enrichment of structures towards the emergence of a system, through a dedicated method called the systemic triple method (DJINDJIAN 1980): 1. Definition of the system S; 2. Perception and description of intrinsic data I; 3. Recording of extrinsic data E; 4. Formalisation of the structuring process: intrinsic structuring (matrix artefact x intrinsic data, O x I), extrinsic structuring (occurrence or Burt matrix intrinsic data x extrinsic data, I x E); 5. Exploratory data analysis on O x I or I x E; 6. Retroactions on I and E; 7. Iterative enrichment by integration of new I and new E; 8. Validation (using another artefact system, a new E, etc.). The modelling level is then examined with a discussion of the limits of the formal logic in Archaeology: empirical-inductive, where 'every structure is Culture', or hypothetical-deductive methods, where 'all the models are fitting well' falling in the weakness of so-called paradigmatic models. A new more restricting method is proposed, called the cognitive model method, CMM. The main features of CMM are: explicit, formalised, repetitive, stable, systemic, refutable, predictive, discursive and auditable. A general method to build a cognitive model is then given, in ten steps; some of them are already known and referenced, others are new and detailed: 1. Improving the knowledge A; 2. Discovering the structures S inside data; 3. Enrichment of structures S; 4. Systemic organisation in hierarchical subsystems; 5. Building models R; 6. Validating models R; 7. Retroactions for enrichment and stabilisation of the models R; 8. Model simulation for predictions; 9. Writing the archaeological discourse; 10. Auditing the discourse. The systemic organisation in hierarchical sub-systems is based on a five level system: 1. Technological know-how; 2. Economic activities: craft production, raw material supplying, subsistence resources, energy resources, buildings (dwellings, infrastructure), territory management, manufacturing, exchange and trading, etc.; 3. Social organisation: workflow, specialisation of professions, social groups, social hierarchies, family structures, community administration, defence, taxes, authority systems, etc.; 4. Symbolic sub-system: ideas and beliefs; 5. Global system. In conclusion, such approaches of methodological development are the most reliable but also the most difficult way to reach a real scientific status for Archaeology.

Cultural resource management (CRM) work in the United States has recently produced vast amounts of data that are now being assembled in large databases. Thus, the potential has grown for useful site location models in support of heritage conservation. As Geographic Information Systems (GIS) have become more powerful, they have become more useful to archaeologists. The realm of archaeological predictive modeling has grown to include at least three types of models that focus either on site-prospection, on understanding ancient ways of life, or on decision-support for cultural resource managers. Decision-support modelling seems to have the greatest near-term potential as a useful modelling tool. However, there are also significant methodological and theoretical issues yet to be resolved before such tools can be widely used. An example of an archaeological site location model currently in development illustrates the potential of decision-support modeling. Some of the problems inherent in site-prospection and ancient-behavior analysis can be avoided in models designed as decision-support tools.

In 1997, within the framework of the research activity carried out by the Istituto per l'Archeologia Etrusco-Italica, a census of GIS applications in archaeology was promoted and then published in the ninth issue of the journal 'Archeologia e Calcolatori'. This international survey allowed us to collect significant methodological and technical information useful to outline the main guidelines of the 'Caere Project', aimed at the establishment of an Archaeological Information System for the analysis of the Etruscan town and its surrounding territory. A recent update of this survey confirmed the previsions published four years ago by the members of an international Scientific Committee coordinated by François Djindjian. Some methodological aspects of the 'Caere Project' are also discussed, with particular emphasis on the need of integrating many different computer techniques in order to gain a comprehensive, organic knowledge of the ancient landscape and town organisation. The results of the innovative use of markup languages and multimedia systems in the processing of archaeological excavation data are also presented. In fact, the methodological approach of the 'Caere Project' is based on the principle that information is not sufficient if not linked by new forms of knowledge representation, which can promote an interactive consultation more than just a passive reading.

A contribution to archaeological resource management. Ghent University has organised an archaeological aerial survey of both provinces of East- and West-Flanders since the beginning of the 1980s. As a result of these activities, some 50,000 photographs have been captured. They reveal thousands of archaeological structures, from the Neolithic through to the most recent periods. Since 1997, financial support has been received from the Flemish Community aimed at the realisation of a GIS based database (Access 97 relational database - Arcview 3.1) and the digitalization of some 50% of the photographs. As a result, it was possible to locate all 50,000 images and connect them with geographical information offered by the support centre GIS Flanders. It is expected in the near future that this information will be available for SMR-purposes and archaeological heritage management. There are also several scientific outputs: one of them is the study of Bronze Age barrows.

The research project, 'Building a cultural landscape model of Minoan peak sanctuaries through a GIS approach', based on a collaboration between the Institute of Mediterranean Studies (F.O.R.T.H.) and the Université Catholique de Louvain, aims to redefine the peak sanctuary, to clarify its function, and examine the relationship between the cultural and natural variables, which characterize the distribution of these sites in the Cretan landscape. To accomplish these goals we used advanced mapping techniques, satellite remote sensing, statistical analysis and Geographical Information Systems (GIS). Intervisibility was investigated with viewshed analysis. A chronological evolution of the peak sanctuary landscape is proposed, explaining the location of the sanctuaries, in relationship to each other and other site types.

Over the last five years global positioning systems (GPS) and electronic total stations (ETS) have become viable tools for use in archaeological field mapping. When used in conjunction GPS and ETS can generate precise, accurate, and georeferenced three-dimensional digital data sets in real time. As survey work proceeds, associated attribute tables incorporating field measurements and commentary can also be created, and the entire dataset can be imported directly into a geographic information system (GIS). This technique may be called precision digital mapping, and produces accurate, high density data sets of unprecedented richness. The revolutions in data management, visualization, and analysis made possible by GIS are now being mirrored by a revolution in field mapping techniques.

Archaeologists working in cultural resource management (CRM) face four major challenges. First, the work is performed under short, development-driven deadlines. Second, CRM projects often are quite large, generating more data than archaeologists traditionally encounter. Third, the results must be presented in forms that are understandable and enlighten the public. Finally, CRM archaeology must address concerns of native peoples. In this paper, I draw on case studies from the American Southwest to show how archaeologists at Statistical Research, Inc. have used the Internet and CD-ROM technology to address these challenges. I highlight a web-based system developed for a large excavation project designed to keep project sponsors, principals, and specialists from around the world abreast of the status of fieldwork and analyses, as well as a forum for dialogue. I also discuss the use of CD-ROM technology to disseminate project materials and reports in a cost-effective manner. Beyond text, these CDs include videos designed for a non professional audience. CD-ROM technology also is used to provide Native Americans with digital images of rock art and sacred sites that allows them to comment on proposed mitigation measures.

A generally Bayesian attitude toward statistical inference seems to me so obviously superior to the 'classical' Neyman-Pearson approach that it is difficult to comprehend why not everyone agrees. I believe that most non-statisticians learn classical procedures ritualistically but then interpret their results in naively Bayesian ways. It would be better if they became more sophisticated and knowing Bayesians. A truly introductory text on the logic of Bayesian inference, with some simple but useful applications, would probably help. Bayesian inference with an uninformative prior may yield the same results as classical inference, but with coherent rather than muddled logic. An example of a very useful but mathematically simple archaeological application of an informative prior is using prior information to improve estimates of true proportions of artifact categories in populations represented by small collections. However, a complication arises when the observed proportion in a fairly large sample is well outside the range considered at all likely for the relevant population, based on prior information. In this case, straightforward use of a beta prior distribution can yield results that seem unreasonable. Possibly our prior information is better represented by a modified beta distribution with 'heavy' tails. Advice about this problem would be appreciated.

Examination of dental development is considered to be an accurate method of ageing non-adults, but ageing adults from dental wear is much less accurate. Miles' method is generally accepted to be the best way we have to derive estimates of tooth-wear ages because it takes into account population variability in wear-rates. Here we develop a Bayesian approach to ageing from dental development and tooth-wear, using a latent trait model and logistic regression to estimate the ages of individuals whose tooth development and/or wear has been scored on ordinal scales. In addition to the original methods this: (a) accounts for uncertainties in tooth development; (b) incorporates in a natural fashion individuals with teeth missing post-mortem. Numerical integrations were performed using Markov-Chain Monte-Carlo techniques and WinBUGS software.

An important field of luminescence dating is the authenticity of ceramic art objects. The use of 'authenticity' instead of 'dating' is due to the ignorance of the ambient radiation, and hence the annual dose. The present paper shows a Bayesian approach able to quantify the degree of authenticity. This approach permits to introduce under mathematical models some assumptions (annual dose, fire, artificial irradiation) previously only presented as qualitative warnings in authenticity reports.

It is well known that interpretation always conveys a certain degree of subjectivity, which disappears as soon as interpreted data are stored in a computer database. This may lead to dangerous approximation and possibly to fallacious conclusions. To avoid this oversimplification, it has been suggested to use fuzzy databases, in which attributes may have a fuzzy nature and be indexed by a numerical coefficient, the fuzzy coefficient, which can be interpreted as the degree of confidence the researcher has in each possible assigned value. This technique has been successfully applied to gender and age assignment for the deceased in a cemetery investigation: in this case anthropological data offered statistical parameters that could be used to compute the fuzzy coefficient. Lithics classification is another field in which fuzzy databases have a potential usefulness, but in this case, no previous statistics may help in determining the fuzzy coefficient. We decided to perform an experiment during a standard typological classification of a flint tool assemblage from Israel. It concerned the classification of 50 tools, by different researchers. Each one was asked to note, besides the typology of each item, an evaluation of the 'degree of sureness', or the 'possibility' of an item to belong to a particular type, in other words his or her guessed estimate of the fuzzy coefficient. This paper reports the results of this experiment, in order to evaluate the difference between researchers when performing a classification of tools, to recognize problematic types or items (which mostly differed between the typological lists presented) and eventually to compute a fuzzy coefficient for each type assignment, balancing the different evaluations of experts.

The archaeological record can be described using a relevant observable feature: location. Shape, size and other properties vary from one location to another, and sometimes this variation has some appearance of continuity, which should be understood as variation between social actions due to neighbourhood relationships. Time and space are not different ways of considering the nature of archaeological locations. Consequently, 'locations' can only be understood in functional terms, that is, according to what is performed at each place at each moment. In this paper, the objective is to analyse where, when and why a social action varies from one location (temporal-spatial) to another. Some mathematical techniques are presented to calculate the probability of social actions at specific locations, based on the spatial properties of archaeological data. These techniques are used as a representation language for studying the concepts of accumulation and attraction, which allow the study of social space in dynamic terms.

This paper presents an alternative to the purely sequential and arbitrary resolution of events in agent-based simulation for Anthropology. It is argued that an alternative system in which agents constantly evaluate their priorities in the light of the actions of other agents provides for more realistic social interaction and allows for the emergence of social-like processes in a computer agent population. A number of other problems, such as the production of survivorship curves from a threat environment, are also discussed.

A new dynamic spatio-temporal model of North American prehistory and protohistory from 14,000 BP to 200 BP allows researchers to visualize the ebb and flow of culture change and demographic processes at any of many possible scales. The authors of past syntheses of such changes over time and space on a large scale in North America have depended upon aggregating lower-level syntheses and summaries prepared by various regional specialists. One advantage of the model is that it eliminates much of the bias and filtering that is typically entailed by this dependence. It does so by directly referencing site-specific data recorded and maintained in a GIS format. These are called up and displayed as animations of spatial change over time. The animations in turn can be mapped against environmental changes over time and space. The model raises theoretical and methodological questions about how we record and disseminate our data. These are briefly discussed.

The results of several sea- and ice-cores and pollen sequences of peat bogs, for the last 40,000 years, permits today to give evidence of palaeoclimatic oscillations of this period. A method of palaeoenvironment reconstitution, based on transfert functions computation, issued from pollen diagrams, is proposed. The method is building a palaeotemperature curve and a palaeohumidity curve, allowing to separate and correlate the two climatic components. A complete example of the method is developed with the data of the peat bog of Tenaghi-Philippon (Macedonia, Greece). All the analysed sequences (Tenaghi-Philippon, la Grande Pile, Banyoles, etc.), compared to sea-cores (KET 8004) and ice-cores (GRIPSummit, Greenland) confirm the evidence of mild and humid oscillations, and cold and humid oscillations, in a three parts structured sequence: the interpleniglacial (up to 28,000 BP), the late pleniglacial (between 28,000 BP and 13,500 BP), and the tardiglacial (between 13,500 BP and 10,000 BP). A numbering system, avoiding usage of ancient interstadials still to valid in their eponym sites, is proposed.