Free text versus chance

Last night I attended the monthly Archives for London seminar. It was entitled “Serendipity in the Archives”. Advertised as an “indulgent evening”, it promised to amuse or even shock. I’m happy to say it was an evening of fun presentations and accomplished all it set out to do.

The idea for the seminar grew from the experience we’ve all had of finding weird or unexpected things in the archives. Usually this is a pleasant experience, made more so when the items found can be shared with others. The speakers gave a lot of thought to the meaning of the term “serendipity”. Several ideas emerged.

One speaker looked at what serendipity means to cataloguing practices and theory. Does free text searching remove the possibility of ‘chance finds’? Does free text searching render hierarchy and sequential referencing obsolete? I think the answer to both these questions is ‘no’.

Free text searches are usually carried out on the catalogues, not the records themselves, whereas items found by chance are usually found in the records, not the catalogue description. Words within an electronic or hard copy catalogue are findable, free text searching just makes this easier. If you’re able to do a free text search of the actual records (isn’t OCR great?), then you will only retrieve the words you look for. If this provides you with something unexpected, then isn’t that ‘chance’?

Perhaps the speaker didn’t take her argument far enough. From a cataloguer’s point of view, does free text searching negate the need for indexing? Does it limit the importance of the title / scope and content division? Indexing archival material is notoriously difficult. A good description will include terms likely to be used as search terms, which arguably makes (subject) indexing obsolete. Many archivists ascribe to this.

What about hierarchy and sequential referencing? Free text searching can liberate researchers from the need to navigate manual findings aids. But, the hierarchy will be more or less important depending on what type of research you’re doing. It’s also important from the perspective of our professional respect for original order. I don’t think we want to get away from hierarchies, although it’s interesting to re-evaluate multi-level descriptions in an electronic environment and to ask questions such as ‘how many levels are too many?’.

Sequential referencing may appear to be made obsolete by free text searching. But its value as a management tool can’t be overlooked. Electronic systems like unique identifiers, which are usually sequential. As human beings, we tend to also like codes that we can make sense of and sequential systems fit the bill.