After our recent meeting in Zuni, T.J. Ferguson outlined a number of different types of information that would be needed to make the database useful to archaeologists. These included better provenience of the objects including the site of origin, its more detailed location within the excavation site, even down to the feature number. T.J. also requested relationships between objects in the sense of them being found in the same feature, grave or layer. He also thought that any information about date (absolute or relative) would be necessary to make the database useful to archaeologists (T.J. Ferguson, pers comm).
Being a trained archaeologist myself, though now a sociologist of technology and information scientist, I understand the importance of these types of information for archaeologists working with collections data. Of course, the fuller context is important to anyone using the information, especially Zunis, to avoid misconceptions or to recognise mistakes. However, I think that T.J.’s request highlights a fundamental misunderstanding of what we are doing with Collaborative Catalogs — that is the difference between a Constructed Catalog and a Collaborative Catalog.
It is not my intention, of course, to single T.J. out on this, as I do not think that he “misunderstands” the system in this sense. His points are perfectly valid in respect to the usual process of constructing a shared catalog. However, his comments brought home to me, very clearly, how easily a confusion may arise between the usual approach of constructing shared catalogs, and our very different approach of creating emergent, collaborative catalogs. So it is not that T.J.’s requests are “wrong”, but that they show how we, as a project, have failed to clearly define what we are doing differently. So, I hope that no one minds, especially T.J., if I use this blog to have a go at making things clearer.
T.J.’s points about including important information about archaeological context is valid generally, of course. The more contextual information available the better for all who wish to use that information. However, the inclusion of common fields, a common ontology, for the shared information points to a model of constructed, shared databases. This model of sharing information has been the norm for many decades, especially when sharing information between associated disciplines and institutions, such as archaeologist and museums. In this model, two, or more, interest or specialist groups get together and thrash out a common data model which will include all the information that the different partners need. This should include categories of information that is needed, but may not yet be available. It may even include information that could be seen to be needed in the future. The mode requires a single common ontology, or information categorisation, that accommodates all the participating partners’ needs, as far as is possible. After such a model is worked out, to everyone’s satisfaction, it is implemented as a single platform for all the access and use.
Today, this mode is often realised as a portal (see Why we are not interested in Portals and Why we are not interested in Portals – 2), the best known for our purposes here being the Reciprocal Research Network (RRN). The RRN has followed this mode to the letter, working with the First Nations in British Columbia to work out a common information model, then creating a shared portal for all to work with and share. Again, it is not my intention to single out the RRN, but, being an innovative and highly commendable project, and one that is setting high standards for such work, it offers an excellent example of this mode of constructing shared catalogs.
The problem with this mode is, and has always been, that it is highly circumscribed, reductive and commensurating. It is circumscribed in the sense that it will necessarily limit the parties who can participate in its construction. Collections are of interest to a vast number, and diversity, of interest and expert groups. Each constructed, shared catalog can only accommodate a small sample of these interests. It is reductive in the sense that even though it is circumscribed to a single or small number of interest groups, it has to create a single model to accommodate both the collecting institution, and the interest group(s). It also has to reduce the diverse needs that always exist within an interest group to a single model, thus being at lease doubly reductive. Finally, it is commensurating in the sense that it assumes that the meaningful identity of the object, the information that is used to describe the object, can be commensurate across different interests, intentions and uses.
Creating Collaborative Catalogs (CCC) is a project that explicitly seeks to overcome these limitations of a constructed, shared catalog for the benefit of all the groups interested in using collections for research and study. Therefore, what we are explicitly not doing is creating a constructed, shared catalog of any sort.
So what is it that we are doing? First of all, to overcome the practical problem of providing information for all of the possible interests around a collection, CCC has devised a whole new model of working with collections information. Drawing on developments in social computing (Boast, Bravo and Srinivasan, 2007), PuSH technologies (PuSH Technology) and emergent systems theory (Turnbull, 2007), Collaborative Catalogs seeks to overcome the problems of centralised databases by distributing both the information and its systemisation to the interest groups. In this mode, rather than sitting down and working out a common, and much reduced, information model, the collections data is given to those individuals or groups who need to use it, and it is they who do the systemisation locally.
So, to go back to T.J. Ferguson’s request, CCC would say, first, that collecting institutions should provide all information that they currently have, but that. second, rather than asking the collecting institutions to be the information hub — which is often beyond their mandate or resources to accommodate, we will PuSH the information to archaeologists and ask them to extract or add this information in a way that systematises and/or extends it for archaeologists. Then, we ask, but do not demand, that the archaeologists offer up this locally systematised and extended information publicly for others to use, including the original collecting institution (if they so desire). Of course, this is not a task just for archaeologists. The CCC makes it possible for all to acquire the information and to systematise and/or extend as is appropriate for them and their knowledge community.
In creating a distributed Collaborative Catalog, in this way, we not only put the information directly in the hands of the expert and interest communities, but we send the information into that community, into its local systems, into the technical and research contexts that can best systematise and extend it for their own needs and uses. This overcomes the three problems with constructed, shared databases in that CCC does not:
- circumscribe the communities that can make use of the information, but diversify them;
- reduce the information set needed by these communities, but accommodating their extension within many local contexts;
- commensurate the information and its use, but to allow for information and use to emerge from a diversity of interest communities.
Such an approach will not only, we believe, engender much more extensive use of collections and their information, but it will create an environment that is more in keeping with contemporary research environments. Environments where many new ways of contextualising and interpreting collections emerge from a diversity of communities and are shared as local solutions to shared problems for others to extend and develop further. In this way CCC is a both a radical departure from conventional data sharing practices, but is much more consistent with developing modes of transdisciplinary research in the sciences (Cameron and Mengler, 2009; Wicksona, Carewc and Russell, 2006; Biagioli, 2009). Like these emerging transdisciplinary research modes, CCC also seeks to vastly extend the expert communities that participate in problem solutions to include source communities and other public stakeholders.
So to finally come back around to T.J.’s request, the answer is yes, but these categories should emerge through an ongoing dialog and co-development by you within your own local archaeological information systems, those of the collecting institutions and the other expert communities. Only in this way, we argue, can we actually achieve the need for local expert systemisation on the one hand, and broad transdisciplinary information sharing on the other.
