This Special Issue of Levant presents a series of papers that represent an outgrowth from a CRANE-sponsored Workshop on the Ceramic Petrography of the Northern Levant, convened at the Department of Archaeology Durham University in August 2015. This particular paper is intended to provide a brief introduction to the CRANE Project to help readers understand the context from within which these papers emerged.

The rapid proliferation of digital data in Near Eastern Archaeology, precipitated by an ever expanding array of data capture technology, has created an urgent need to establish a collaborative research environment with the capacity to address long-standing issues of access, data compatibility, integration and analytical capability. The Computational Research on the Ancient Near East (CRANE) initiative is a multi-disciplinary consortium, made up of archaeologists, historians, palaeoenvironmentalists and computer scientists, that seeks to facilitate the creation of such a collaborative scholarly framework. CRANE has focused initially on the Orontes Watershed, a cohesive geographical unit and region, uniquely positioned as a cultural microcosm of the broader ancient Near East. It was thus viewed as an ideal operational test case.

CRANE aims

As an international, collaborative Near Eastern archaeological research initiative, CRANE seeks to achieve the following research aims:

• assemble digital data from multiple archaeological projects that use different terminologies and recording systems, an outcome of the diverse research traditions practiced in the East Mediterranean;

• develop a fully integrated, temporally and spatially controlled cultural and palaeoenvironmental record for the Eastern Mediterranean that can support complex analyses of heterogeneous datasets drawn from this record in both time and space;

• model forward and inverse simulations of past human-environment dynamics and a range of social practices, including both climatic and anthropogenic impacts, based on parameters supplied by empirical datasets;

• create spatially accurate and realistic 3D visualizations of reconstructed ancient landscapes and human activity based on empirical data and the output of simulated scenarios;

• develop protocols and analytical tools to facilitate broad access to this information; and

• create unique research opportunities and training in advanced archaeological analysis for students (BA, MA and PhD).

The CRANE collaborative research network is not an end in itself, but a means to achieve a deeper understanding of the foundational societal and cultural developments in the ancient Near East that have shaped the cultures and landscapes of the present; especially in the modern Middle East. We believe that a deeper and more nuanced understanding of past trajectories will provide direct insight into such contemporary concerns as the continuing ecological impact of ancient anthropogenic changes to the Middle Eastern landscape (e.g., deforestation, soil erosion and salinization); the need to employ more effective adaptive strategies in marginal environments, including sustainable agricultural and horticultural practices; the long-term health consequences of traditional dietary practices and subsistence strategies, especially in urban communities; the socio-economic and political impacts of climate change; and the cultural roots of long-standing political conflicts, and their continuing impact on current events.

The OCHRE system and archaeological data integration

Advances in computing technology have enabled the increasingly rapid processing of ever-expanding databases in the humanities and social sciences. However, the quality of this digital information is dependent on its prior organization, and on the corresponding software techniques employed. The rapid increase in the quantity and diversity of digitized data now available, particularly in the humanities, has exposed the limitations inherent in the simplistic database structures and underlying data models typically employed (McCarty 2005). Research has been further impeded by the fact that data compiled by a single scholar, or group of collaborators, are not easily combined with data compiled by other scholars in a way that preserves all of the idiosyncratic analytical distinctions and interpretations contributed by the various creators of those data.

This is a particularly pressing problem in archaeological research, with its reliance on heterogeneous datasets; in recent years leading archaeologists have made a cogent appeal for a solution to this problem (Kintigh 2006; McManamon and Kintigh 2010; Snow et al. 2006). A number of attempts have tried to impose uniform data-recording systems and descriptive taxonomies, such as the ETANA-DL (Ravindranathan et al. 2004) and ARCANE (Lebeau 2014). However, idiosyncratic description and interpretation should not be seen as a vice to be corrected, but rather as a defining characteristic of the humanistic mode of research, and thus these attempts to impose standardized taxonomies have largely failed. Large-scale data analysis has, nevertheless, become increasingly common in humanities scholarship, and archaeology in particular, although all too often with impoverished results, reflecting the difficulties inherent in machine processing of natural language (Dreyfus 1992), and in the integration of heterogeneous datasets that conform to divergent and varyingly structured schemas (Doan et al. 2012; Garcia-Molina et al. 2009). These very issues have traditionally hindered the comparison of ceramic data between sites and regions.

The challenge of integrating heterogeneous datasets comes into sharp focus in the discipline of Near Eastern Archaeology, which incorporates the study of diverse textual sources with non-textual material cultural remains. Near Eastern Archaeology thus represents an ideal test case for developing a computational framework capable of integrating heterogeneous datasets into a single query-able environment without loss of information (Kansa 2005; 2010; Petrovic et al. 2011). The key operational features are flexibility and extensibility, with a core ontology in which as few concepts and relationships as possible are predefined, enabling a high level of abstraction and reusability in the software that implements the data-integration system.

Towards this end, CRANE has embarked on a partnership with the Online Cultural and Historical Research Environment (OCHRE) at the University of Chicago (Schloen and Schloen 2012; 2014). OCHRE is a computational platform whose core feature is its flexibility and highly atomized data structure. OCHRE’s ontology is inherently extensible, permitting researchers to construct their own lists and hierarchies of data items, and their own taxonomies to describe those items. OCHRE shares similarities with another recent archaeological computational initiative, the Digital Archaeological Record (tDAR, see www.tdar.org; also McManamon and Kintigh 2010; McManamon et al. 2010), which attempts to integrate archaeological datasets by correlating disparate ontologies, or taxonomies, on-the-fly. OCHRE, by contrast, integrates heterogeneous datasets within a single enterprise-class database system by means of a global schema that prescribes the logical structure of all the data and reflects OCHRE’s core ontology. This flexible structure permits an efficient and user-friendly ‘data warehouse’ approach in contrast to the more cumbersome ‘data mediation’ approach of tDAR. OCHRE is especially useful for research groups working on related material or research problems, allowing them to integrate their data rigorously and conduct detailed cross-project comparisons and analyses. Moreover, OCHRE makes it easy for such collaborations to grow and become interconnected as research progresses, providing a mechanism for wider-scale data integration from the bottom up as new projects join and new thesauri are created to match terms from different taxonomies.

The CRANE data integration effort is still in its early stages, but has already achieved significant progress. Thus far more than three terabytes of digital data have been fully integrated into the CRANE computational system, allowing for many millions of relational links. The data are comprised of a remarkably diverse range of formats, including image files (photos, maps, scans of drawings, notes; formatted as JPG, TIFF, PDFs), vector (formatted as SVG files), GIS and 3D data (formatted as SHP, GRD, DEM, DXF, KMZ, OBJ, SKF, FBX files), climatological data, text, numeric and tabular data (formatted as DOC, RTF, TXT, CSV, XLS files), linguistic data (including transcriptions, transliterations and translations of ancient texts), bibliographic data, and XML metadata.

CRANE also seeks to create computation tools that will facilitate the modelling and visualization of the inter-relationships of social, economic and environmental dynamics at multiple spatial and temporal scales of analysis in order to gain more meaningful insight into the rise and development of complex societies in the ancient Near East. To facilitate these broader research objectives, CRANE has initiated a number of data gathering sub-projects. These include (1) the CRANE Regional Site Database, a comprehensive site inventory encompassing the entire Orontes Watershed, currently totalling 4414 archaeological sites dating from the Neolithic to the Ottoman period; (2) the Ceramic Repository of the Orontes Watershed (CROW), a ceramic database and shape-matching project; (3) the creation of a robust absolute timescale for the Orontes Watershed calibrated to the decadal level, involving the wiggle-matching of a matrix of radiocarbon dates from multiple sites in the watershed with a regional dendrochronological sequence; and (4) the compilation of palaeoenvironmental data, both from sediment cores and historical records that contain information about the climate, geomorphology, hydrology and vegetation of the region (for further details on these sub-projects, see www.crane.utoronto.ca).

Ongoing and future directions

The ongoing cultural heritage crisis in the Middle East has created a further urgency for better collaboration and the sharing of data about culturally significant sites and objects. CRANE was launched as an international research collaboration prior to the outbreak of the present conflicts in the region, and thus its research aims and objectives were not determined by the cultural heritage crisis that has since ensued. A wide range of potential stakeholders nevertheless were envisioned when the project was conceived, including both governmental and non-governmental cultural heritage management agencies and organizations. The present crisis therefore also represents an opportunity for the Near Eastern archaeological community to respond meaningfully to issues of pressing current concern.

There are also more pragmatic reasons to collaborate and share data. These include — perhaps counter-intuitively — the need for ever-stronger security protocols for protecting data, and the general global trend toward greater access to information. Regarding the latter, it is worth noting that national research councils are moving steadily toward mandated policies that require grant recipients to make their research results publically accessible, on average allowing for a one-year embargo period. The Canadian Research Councils, for example, recently initiated such a requirement for all grants awarded as of 1 May 2015. Consequently, it is safe to assume that recipients of public funding will increasingly find themselves restricted to two knowledge mobilization options: (1) the deposit of their research data into open access repositories, and/or (2) the publication of their research in open access journals, whether they wish to or not. National antiquities authorities are also moving increasingly toward similar policies. While Open Access remains highly contentious, and the issues complex, this trend will likely only continue. A prudent disciplinary response should, therefore, involve strategies that collectively and collaboratively engage this brave new world.

As I have tried to argue, Near Eastern Archaeology represents an ideal test case for developing a computational framework capable of integrating the complex, heterogeneous, or ‘messy’, datasets that typify our world, and defy the homogenizing algorithms of machine-driving learning. Moreover, the most successful collaborative data-sharing ventures are likely to be those that accommodate multiple, diverse taxonomic structures and hierarchies of knowledge and meaning. It is our hope that the success of this workshop illustrates the analytical capability and research utility of the CRANE collaborative approach.