Maurizio Forte, Arkadiusz Marciniak

Maurizio Forte, Arkadiusz Marciniak

The metaexcavation: A 3D mode of inference at Çatalhöyük and beyond

1.Introduction

The site of Çatalhöyük in Central Anatolia became of international significance because

of its size and complexity at an early date outside the Fertile Crescent since it was first excavated between 1961 and 1965 by the British archaeologist James Mellaart. From 1993 onwards archaeologists from a variety of countries, including teams from the United Kingdom, the United States, Greece, Poland, Turkey, Germany, Italy, Sweden, among other have investigated the site as a part of the large project directed by Ian Hodder.

The Çatalhöyük project is one of the most complex archaeological projects conducted today. It incorporates the newest techniques aimed at studying the site of a highly complex stratigraphy. It integrates a wide of specialists that works hand by hand with the excavators at every field season. The project addresses a wide range of research questions by mobilizing an ever increasing number of datasests. This makes the archaeological process highly heterogeneous and idiosyncratic as it requires a constant compromising of the heuristic potential of a range of different categories of evidence being recovered at the site.

The duration of the project and participation of archaeologists originating from different research traditions resulted in implementing a range of diverse and approaches to the excavation process. Additionally, as the project takes place in a dynamically developing field methods, the Catalhouyk methodologies had to constantly implement different new techniques at different levels of the research process such as digital photos, relational database, GIS or 3D data recording. These were aimed at significantly improve and make the inference process as efficient as possible.

We would argue that despite introducing a wide range of new methods and techniques it is only introduction of a 3D simulation marks a qualitatively new phase of the research process at archaeological sites. It is mainly because it shall facilitate a new mode of inference that shall fundamentally change the archaeological process. In this paper we intend to present major facets of this qualitatively new approach, in particular looking at the ways in which this makes it possible to integrate different datasets. It will start by presenting the state-of-the-art of archaeological process at Catalhoyuk, which is a point of departure and reference for discussing a range of new developments associated with the 3D simulation inference. It will then present a major facet of this theoretical breakthrough, which is benefits of the virtually reversible excavation process in a simulated environment from laptop computers to virtual immersive systems as a means of dealing with these new challenges. The paper will then discuss some potential further developments based upon 3D simulation inference model.

2. History of Catalhoyuk methodologies

Explicit methodology was defined prior to commencement of the fieldwork and prior to arrival of the international teams to handle the project’s objectives and to deal with ‘the challenge of introducing multivocality and reflexivity in the laboratory and trench’, as formulated by I. Hodder. The elements of this new approach comprised: (a) priority tours aimed at the discussions between the laboratory and field staff, (b) interpretive approaches to sampling strategies, (c) co-operation of specialists at the site, (d) quick feedback by the laboratory staff to the field staff, (e) interactive database available on and off the site, (f) the writing of a diary to enhance a fluid and flexible data, (g) video recording, (h) presence of social anthropologists studying the construction of knowledge at the site, and (i) hypertext solution to challenge the linearity of archaeological narrative and allows accounts with multiple pathways and multimedia. (cf. )

The international nature of this large project led to incorporating a range of different tradition into this model. They refer to (a) excavation methods, (b) sampling procedures, (c) recording system of different categories of data, and (d) relationships and co-operation between the field and laboratory staff.

Due to time constraints, we will focus here on excavation methods only. A modified form of single context (unit) excavation and recording was employed and the unit forms became the basic element of a nested hierarchical system. It is generally referred to as a ‘context’ in British field archaeology and ideally represents a single identifiable depositional event. The system also included features (understood as groups of related units), space (spatially bounded entities generally defined by the walls of buildings), buildings (groups of spaces forming a structural entity) and areas (spatially discrete locations where excavation had occurred, as in the South Area) (Hodder 2005, volume 6).Chronological grouping was provided by phases and levels. This model originated in the single context system of excavation and the recording developed in British urban archaeology in the 1970s and now employed as standard practice in contract archaeology in UK (Harris 1979; Harris et al. 1993; Roskams 2001; Spence 1993) (Farid, Cessford).

The initial system developed in 1995 involved five general unit categories: layer, arbitrary layer, cluster, skeleton and cut. In 1997 a further level of ‘interpretation’ was introduced as the general categories did not provide sufficient information to the laboratory teams about what type of deposit was under investigation, especially when information was required at a glance. Initially, excavators used a diverse range of interpretative terms. This allowed for an individualistic and 'fluid' approach, and it led to an exploration of the range of terms deemed necessary for this particular site. As the post-excavation process started in 2000, however, a systematic set of terms was introduced and labelled ‘data category’. The ‘data category’ was introduced to enable a query of the same type of events on the database. A standardised list was compiled on the basis of the deposit types excavated on the site over the few preceding years. It was accepted that new types of deposits could be encountered as excavation progresses and the list would be amended accordingly. There were currently ten primary data categories of fill, floor, construction, midden, activity, natural, arbitrary, cut, cluster and skeleton.

Single-context excavations put in place at Çatalhöyük had some shortcomings as compared the Polish methods. This refers in particular to the practice of cutting the feature into two halves, digging one of them and then drawing its section. Moreover, the category of unit was unknown in the Polish excavation methodology, where the feature comprised of its basic entity. It took us some time to get acquainted with the unit-feature scheme. The Berkeley team’s , are said to stem from the experience in Balkan archaeology. The team leaders re-defined their methods to achieve their aims to study a life history of the house rather than settlement as a whole, which was the case in their work the Balkans (p. 112). This required that all remains are systematically and fully mapped, recorded and sampled to a degree never occurred ‘on any other Neolithic site yet excavated in southeast Europe’ (p. 112). This strategy was also different from investigation of the vertical exposure of the stratigraphic sequence.

In spite of sharing of methodological principles and aims of the Çatalhöyük project, there were some important differences put in motion by both teams. In both the Polish and the American practice, excavated structures were monitored vertically by cross-section, which was not the case at Çatalhöyük. As a result, the Berkeley team retained the use of a cross- section while the Polish team did the same while digging more complex stratigraphic sequences. The American team continued to use small temporary profiles to understand microstratigraphic relations, whenever possible. They aimed to demonstrate and document stratigraphic relationships and supplement the Harris matrix analysis. This approach, however, was not implemented systematically as compared with excavations in the Balkans or Central Europe. Moreover, the Berkeley team advocated a need to excavate by arbitrary layers, a practice known from southeast Europe and the US in the case of thick and undifferentiated layers of mixed materials such as house fills.

There was no doubt, however, that there was a number of significant benefits of this kind of large international enterprise and possibility to look at the way in which archaeology was being practised ‘at the trowel edge’ by various national teams and co-operation with a vast range of specialist at this level of archaeological practice. Moreover, a micro-focus of the project is particularly to be praised.

3. 3D inference process

As mentioned in the beginning, the introduction of 3D mode of recording marks the beginning of qualitatively new phase of the research project at Catalhouyk, This new approach not only facilitate and significantly speeds up the excavation process itself and makes possible to more efficiently integrate different datasets with the excavation data but more importantly marks a completely new and previously unknown inference process.

The elements of new recording system and analysis of excavated strata and structures at a trowel edge comprise (a) quasi-real time digital processing, (b) paperless data entry, and3D visualization. Every phase of excavation whereas the diggers recognize new stratigraphic units (following the single context method) is recorded by laser scanning and computer vision, in this case using standard digital cameras and the software Photoscan. The outcome is a 3D databases constituted by 3D meshes, clouds of points and texture of all the archaeological units. In post-processing it is possible to extrapolate the units from the models according to their boundaries and as they are recorded as single logical entities linked with metadata.

At the moment the experimental project on the B89 has produced 97 3D phases of excavation and hundreds of stratigraphic units. The 3D models are available on site during the excavation and this allows a continuous interaction between empirical (stratigraphy) and mediated data (digital). One of the most important challenges of the simulation virtual process is the review of the entire excavation with the ability to modify and re-interpret the logical sequence of stratigraphy. In other terms, every unit is contextualized within the 3D building together with former excavated units and other archaeological evidences, for instance, features that are not identified yet as unit or former excavated units (i.e. negative units, pits, etc.). In the case of a Neolithic building all these information are visualized in 3D during every single phase of excavation, so that diggers can evaluate the stratigraphy in relations to other structures of the building or surrounding 3D spatial data. The excavation becomes a visual 3D puzzle to investigate and decode.

The particularly challenging is the excavation process itself aimed at systematic recording and recovering these different data using a complete set of recovering, recording and documenting techniques.These new methods make a significant step further while comparing traditional methods of excavation with 3D digging. In fact 3D recording and modeling give the archaeologists a complete and exhaustive view of the excavation, actually contextualizing all the information in the same virtual space and saving all the data in digital format. Geometrical information, textures, databases can therefore generate different sequences or can validate and support more objectively the archaeological interpretation.

This new ability for archaeology to operate even virtually within the archaeological excavation opens new perspectives. Textual forms, databases, photos, models, maps and related information can regenerate a potential new and unexplored digital excavation, where the interpretation is enriched by all the information surrounding the user/archaeologist. This “enriched” excavation stimulates more advanced inferential methods of interpretation and data validation/

To sum up, new elements of the inference process include (a) digital reversibility of the archaeological excavation, and (b) assembling different kinds of data and research questions. This contribution of the 3D model is far the most important facilitating very efficient processing of data, making the process repetitive and reversible.

As a result of these recent developments, Catalhoyuk is now completely papers less and 3D model is omnipresent. This has far reaching consequence that will bring the inference process into hithero unthinkable and non-existing level. .One major development comprises a different procedure of dealing with inevitably fragmented different logical entities (empirical, on paper, digital). They can now be assemblage and re-assemblaged in a potentially endless way facilitating more advanced level of interpretation and knowledge transmission This marks a significant departure from the existing solutions which can be described as linear and in which subsequent datasets are kept added in a cumulative way. The 3D model makes also possible to overcome some limitations of single context recording, much more efficiently than proposed above. Unit is defined here on the basis of non-technological approach creating a gap between excavation and data recording. These two elements appear to be completely separated from each other. The 3D model clearly shows that deposition is much more complex and makes it possible to grasp its nature, which has far reaching consequences for the analysis of a number of variables.

The 3D model makes also possible to be a valuable asset to the now fashionable Bayesian analysis. It is based upon rather traditional understanding of stratigraphy. The method is based upon grouping of layers while we know very little about complexity of these layers. For example the excavation of burials at Çatalhöyük is very challenging since there are always several depositional and post-depositional activities. Every new deposition involves the opening of the platform/ritual pit, skeleton/skulls removal, inclusion of new individuals, final sealing with plaster. This sequence is complex and repetitive but it is very difficult to be recognized during the archaeological excavation: the upper level/skeletons are not necessarly the latest ones and it is complicated to identify any single context and separate burials. The 3D documentation is an extraordinary help, as demonstrated in the B.52 and 89, because it is able to re-interpret and reconstruct the archaeological sequence even after the excavation, suggesting alternative perspectives.

The new approach offers also a good solution to deal with the ever existing challenges of the Catalhoyuk project. As the problem driven excavations it requires a contestant mobilization of different categories of data, These comprise: (a) coming to terms with complex recording protocols, different datasets that require different databases, spatial distribution of different data, codependence between different categories of data; (b) scale of excavations with a large number of teams working side by side

This “meta-excavation” process, conducted within the “3D-Digging at Çatalhöyük” project, makes it possible to merge the reflexive process faced by the archaeologist with empirical data (archaeological evidence) and mediated experience (the virtual simulation). The interaction between empirical data and virtual simulation is a new challenge in archaeology. Given the rapidity of the simulation process, the operator is able to compare for example a 3D record of a stratigraphic unit, with the unit itself. The first is visualized in a tablet or a laptop, the second one of course in a stratigraphic sequence in situ. This new interaction stimulates new research questions and actually augments the potential to re-interpret archaeological evidence. In other terms, the virtual simulation multiplies views, perspectives, hypotheses, beyond the usual instruments of dissemination, such as reports, printed publications, archives and so on…

Actually, there is not an opposition between empirical data and mediated experience, since in any case even the bodily interaction with layers/stratigraphy is somehow a mediated experience. The real difference between empirical data and simulation is in temporality: empirical data can be experienced just during the excavation and after that they are gone, while a virtual simulation can be repeated indefinitely. We argue that the reiteration of experience by digital simulation, increases substantially the informational process and includes a larger number of stakeholders to the realm of interpretation.

4. Summary and outlook

The implementation of 3D model is of revolutionary character for the Catalhoyuk Research Project and paves the way for radical changes in the archaeological process and mode of inference far beyond the Catalhoyuk excavations. In particular it will change the excavation process and inference protocols in relation to the hitherto practiced. It makes it possible to efficiently process an enormous amount of information produced in a season of fieldwork and in a very short time. The accessibility and processing of digital archaeological data can change the interpretation process in the future (i.e. involving a large community in analysis, interpretation and communication of data).

An in-depth understanding of this process is possible thanks to ongoing studies of the relations between human brain and 3D immersive systems. The authors argue that this ‘meta-excavation’ process provides an unprecedented possibility to deal with the challenges of a rapidly changing discipline by providing solid foundations for constructing bold inferential modes using different range of datasets and consequently enriching the archaeological interpretation.

3D archaeology is a new domain and introduces new and more advanced inferential methods of interpretation. This doesn’t mean automatically the achievement of better results because of the persistent ability to use bi-dimensional mental maps in the excavation process. “Thinking” in 3D is something new: new perceptions, awareness, connections, affordances. The migration of 3D worlds in immersive systems, such as “caves”, haptic systems, holographic projections and so on, generates a different kind of embodiment and spatial relations body-environment. This different “3D brain involvement” is not yet adequately studied but it is a new challenge also for neurosciences to understand this approach to alternative realities. An interesting example is the use of Oculus Rift, a new and portable head mounted display. This system uses an omnidirectional sensor and allows a very accurate perception of the scale and spatial presence in the virtual environment. This augmented embodiment is able to stimulate additional affordances and a deeper sense of tangibility of digital objects.

In conclusion, the metaexcavation represents a new frontier of virtual simulation and it faces new issues in understanding the complexity of 3D connections and visualizations. In addition this approach foster to review also the single context excavation, in the sense that the 3D holistic approach to the excavation highlights different kind of relations among layers, strata, finds and object, not accessible or visible with a traditional approach.

1