Mediterranean archaeology and the Not-So-Open Sea
Just a quick contribution, following up on an email discussion that Stefano Costa and I had about current attitudes towards, and provisions for, open access in Mediterranean archaeology (and reminded by some conversation at a fun conference a couple of weeks ago in London).
I think some interesting discussion could be had about the intended detail of data-sharing amongst those involved in Mediterranean archaeology, or indeed archaeology more generally. For example, there are already some very useful online summaries of Mediterranean excavations and surveys in the form of such things as AIAC and L-P Archaeology’s FastiOnline or Tel Aviv/USC/UCLA and IPAWG’s West Bank and East Jerusalem Archaeological Database both of which I think are the kinds of ambitious, but effective cross-border or cross-party efforts that hopefully will be taken up more widely. A great example of a Mediterranean typological resource is the Southampton Roman Amphora project. Another useful resource on Mediterranean surface surveys is CGMA’s MAGIS site, albeit with restrictions on usage and no immediate way to download raw data or results of searches at present (please correct me if I am wrong).
In any case, this is of course far from an exhaustive list and reflects my own research biases. Such Mediterranean projects have a variety of valid priorities and working constraints which lead to the particular dissemination approaches they have chosen. Not everyone will of course agree with all of them and there is, I’d argue, still some further championing of direct-access-to-simple-text-and-image-files to be done. More to the point, the acknowledged focus of these initiatives is on coarse locations (e.g. site centroids), typologies (e.g. of representative artefacts only), and metadata (general context of research, index level bibliography). Such standardising, integrative efforts are clearly crucial, but it would be a pity if that was the only sense in which Mediterranean or any other regional archaeology sought to share data.
In contrast, full, raw and open datasets (for example of excavation and survey archives at the scale of individual recovery units, original finds databases, etc.) are made public only to a very limited degree, even many years after collection. The best ones I have come across so far are typically those driven by the UK Arts and Humanities Research Council’s requirements for archaeological projects with funding from them to archive with the UK Archaeology Data Service’s (ADS; e.g. search for “Mediterranean” here). The Open Context initiative also looks promising although still in its early stages. Even so, the degree to which these project archives really do provide usable, primary datasets varies a lot to my knowledge, and the ADS still has a click-to-agree-with-our-terms protocol which tends to get a bit in the way (any comment on the formal rationale behind this?).
In any case, whether such raw data provisions are achieved via integrative mega-projects, official repositories, open access journals with dataset DOIs, or other arrangements (one of the latter I and some colleagues are working on as a companion to an imminent ADS deposition is here: and the Çatalhöyük database is certainly worth a look) and different kinds of funder or permit requirements is less important (though on the mega side of things, there is clearly a case for well-funded, discipline-agnostic national or trans-national depositories: e.g. building on the idea of Public Data Corporations) than the fact that there currently is such a huge missed opportunity. Particularly strange that the Mediterranean, despite being one of the most data-dense archaeological records in the world, provides so little raw data! The straightforward combination of one or more simple ascii data files, a short description explaining some archaeological background and an open license all have a beautiful simplicity to them — we should start with that, rather than getting to0 worried about either semantically-rich-and-structured stuff or a full theoretical barrage about ontologies etc.
While I am at it, one final issue with regard to Mediterranean spatial datasets and full data-sharing is, of course, the risk of promoting some kind of spatially-enhanced looting (and for the issue in archaeology more generally, Ant Beck’s blog is great). This gets especially tricky of course when our open data efforts cross modern political borders (there is a good blog and journal special issue on this)? Such risks are probably less relevant for excavation archives and finds databases of known sites (“unencumbered data” as Stefano puts it), but are certainly a theoretical worry for many who deal in site-level spatial data across whole landscapes and in known areas of looting. In any case, my feeling is that a) this is a fear that, while very plausible in theory, is rarely backed up with sufficient proof in terms of much documented looting behaviour that has been enabled by academic digital publications of spatial coordinates (unless I have missed something – perhaps shipwreck ‘salvage’?), b) such an issue is being rapidly taken out of specialist hands (and therefore over the next decade will effectively become less of an issue, whatever our misgivings) by the fact that non-specialists and informal contributors (site visitors with cameras, GPS etc., locals, metal detectorists, project participants, etc.) can now contribute fairly precise locations of cultural heritage finds and sites to Google Earth etc., and already promote spatially-precise exploratory activities through (what are usually ecologically progressive) hobbies such as geo-caching. Access to Mediterranean datasets with intentionally-degraded spatial information and/or limited to those with a state or academic institutional affiliation, are perhaps the two most often mentioned spatial firewalls in sensitive cases (e.g. the different levels of coordinate access in the UK’s Portable Antiquities Scheme – which is worth a look more generally, if you have not seen it before, as it is simply such an important initiative), but neither are as open as we might otherwise wish. Beyond simply saying that we agree that open-spatial-data-enabling-looters is an important issue and that it is part of an ongoing discussion, how programmatic do we need to be?
Anyway, thoughts and corrections on the above are very welcome!