So here’s the way we do things in the archive world: You have your boxes of papers, all messy and disorganised. Then you try and impose some order on them, trying to make sense of how the records were originally created and maintained. After a well-earned tea break (of course, a safe distance away from the original, unique documents!), you start to organise them into neat and pretty categories and form a lovely tidy hierarchy, which supposedly reflects the structure and organisation of the creating body. Or that’s the theory anyway.
But how do you impose this hierarchical order and structure on a collection when the creating organisation is defiantly non-hierarchical, when the organisation itself has always operated within more of a matrix framework, when the records in question are not the work of one body, but instead, the work of many hands – of all the individuals who make up that institution.
For the Tavistock Institute of Human Relations (TIHR), this is the case; the Institute has only had a Director/CEO role in relatively recent history, and has always organised itself in a lateral rather than hierarchical fashion, allowing its members to work across boundaries rather than within defined silos. Over its history, there have been a number of different units and working groups, as well as a number of individuals collaborating on specific projects at any one time. A hierarchical structure would feel like an over-simplification, or – worse – a falsification, of how the Institute works, and has worked in the past.
I’ve been mulling this over recently, accompanied by many cups of tea to help the brain cogs turn. It’s made me think again about how archivists choose to catalogue records, and question that thing of doing something just because it’s The Way We’ve Always Done Things.
Following the traditional way of doing things (the record group concept), a hierarchical arrangement would describe the creating organisation, followed by a description of the sections, series, and items created and managed within this environment (think of it like a big flow chart, where everything inherits information from above).
Traditional approaches to cataloguing would have me place the TIHR at the top of the hierarchy (or a family tree), and then underneath that creating body there would be a cascading series of records created by the different parts of the Institute, representing its different departments or functions. These nest within each other, so you can see that a particular department sits within a particular division, and so on up the tree until you reach the ultimate creator of the records: the TIHR. So far, so good, but cataloguing in this way does imply a rather rigid neatness, a sense that there is a very straightforward one-to-one relationship between creator and record.
It ignores the messiness and real life nature of records and organisations, in which there is a complex web of dynamic relationships. It ignores the fact that records are created by individuals, who operate within groups, which form and merge and split and remerge, and instead tries to impose an over-simplified order which distorts the reality of how these records came into being.
Faced with the records of TIHR, a non-traditional, complex and dynamic organisation, I needed to find another approach to cataloguing. Something which could embrace the messiness and interrelatedness of records and their creators, rather than trying to pigeon-hole things into clearly defined taxonomies. This is when I started considering using a series system approach for cataloguing the archive.
The concept of a series system was originally developed by Peter Scott in Australia, to document how environments of complex administrative change impact on records creation and management. His approach was to separate descriptions of the authorities (eg, persons or organisations) who generated records from the record descriptions themselves. In this way, you can create more of a matrix which maps out the interconnected relationships between people, activities and things; provenance is expressed by linking records to record creators, allowing multiple and complex provenances to be established through linkages.
The architecture of the series system is based around linking authorities to records in order to establish context, meaning more than one individual or group can be linked to any one record. It also means that questions over where to place records within a hierarchy are made easier; as papers can almost be effectively in two places at once.
For example, let’s say I’m cataloguing papers created by Eric Miller, relating to the Calico Textile Mills project, which Miller worked on, but also a number of other individuals too. In a traditional hierarchical structure, a decision would need to be made about whether to classify these records as either the records of the Calico Textile Mills, or the records of Eric Miller. However, using the series system, it is possible to tag the series so that it is connected to both Eric Miller (as a name authority) and the project itself, as well as to any other individuals involved in the project. This sort of complex reality is more accurately expressed by the series system than by the record group system.
This offers up a whole new range of opportunities for researchers. It means that you don’t need to navigate through a static catalogue to find the records you need, or rely on an (often problematic) key word search to find what you’re after. Instead, you could search for the records of the Calico Textile Mills, for example, then see that Eric Miller is tagged as an authority – click on his name and you’ll bring up all the records of projects Miller was involved in. From there you might see that Miller worked as part of a particular working group or unit, and you could click to see which projects that unit worked on. So it would (hopefully!) create a more dynamic way of searching across the catalogue, in a more lateral and accessible way, recognising that researchers may be interested in accessing the collection for a variety of different reasons.
Something of the newness, the challenge, and the novelty of series cataloguing seems to mirror the working of the Institute. This is a cataloguing method which, although widely used in Australasia, has never really caught on over here, other than at a few archives. It’s certainly a first for the Wellcome Library – with only one other project (Wellcome Trust’s Corporate Archive project) currently underway which uses this methodology. This means having conversations with those other UK institutions who are using series based approaches (House of Fraser and TfL are two key examples), and trying to learn from their stories and experiences. There are also other implications about how the metadata will present itself, about how to manipulate our cataloguing software so that it does what we want it to, and other technical matters that we’re still scratching our heads over.
There is both anticipation and excitement about the idea of cataloguing in a new and different way; there is a lot that will need to be worked out along the way. In a way, this is what appeals most about adopting a series based approach for TIHR; it is not only a way of more accurately reflecting the organisational design of the Institute, but also a more dynamic and playful way of working with the data, an experimental piece of work which seems to speak to the work of those early TIHR pioneers.