This is (probably) a final blog post referring to the recent survey by the UK Archives Discovery Network (UKAD) Working Group. Here we look in particular at subject indexing.
We received 82 responses to the question asking whether descriptions are indexed by subject. Most (42) do so, and follow recognised rules (UKAT, Unesco, LCSH, etc.). A significant proportion (29) index using in-house rules and some do not index by subject (18). Comments on this question indicated that in-house rules often supplement recognised standards, sometimes providing specialised terms where standards are too general (although I wonder whether these respondents have looked at Library of Congress headings, which are sometimes really quite satisfyingly specific, from the behaviour of the great blue heron to the history of music criticism in 20th century Bavaria).
Reasons given for subject indexing include:
- it is good practice
- it is essential for resource discovery
- users find it easier than full-text searching
- it gives people an indication of the subject strengths of collections
- it imposes consistency
- it is essential for browsing (for users who prefer to navigate in this way)
- it brings together references to specific events
- it brings out subjects not made explicit in keyword searching
- it enables people to find out about things and about concepts
- it may provide a means to find out about a collection where it is not yet fully described
- it maximises the utility of the catalogues
- it helps users identify the most relevant sources
- it can indicate useful material that may not otherwise be found
- it enables themes to be drawn out that may be missed by free-text searching
- it can aid teachers
- it helps with answering enquiries
- it facilitate access across the library and archive
- it meets the needs of academic researchers
The lack of staff resources was a significant reason given where subject searching was not undertaken. Several respondents did not consider it to be necessary. Reasons given for this were:
- the scope of the archive is tightly defined so subject indexing is less important
- the benefits are not clear
- the lack of a thesaurus that is specific enough to meet needs
- a management decision that it is ‘faddy’
- the collections are too extensive
- the cataloguing backlog is the priority
Name indexing is considered more important than subject indexing only by a small margin, and some respondents did emphasise that they index by name but not by subject. Comments here included the observation that subject indexing is more problematic because it is more subjective, that subjects may more easily be pulled out via automated means and that it depends upon the particular archive (collection). As with name and place indexing, subject indexing happens at all levels of description, and not predominantly at collection-level. Comments suggest that subjects are only added at lower-levels if appropriate (and not appropriate to collection-level).
For subjects, the survey asked how many terms are on average applied to each record. According to the options we gave, the vast majority use between one and six. However, some respondents commented that it varies widely, and one said that they might use a few thousand for a directory, which seems a little generous (possibly there is a misunderstanding here?)
Sources used for subjects included the usual thesauri, with UKAT coming out strongest, followed by Unesco and Library of Congress. A few respondents also referred to the Getty Art and Architecture Thesaurus. However, as with other indexes, in-house lists and a combination approach also proved common. It was pointed out in one comment that in-house lists should not be seen as lesser sources; one respondent has sold their thesaurus to other local archives. There were two comments about UKAT not being maintained, and hopes that the UKAD Network might take this on. And, indeed, when asked about the choice of sources used for subject indexing, UKAT again came up as a good thesaurus in need of maintenance.
Reasons given for the diverse choice of sources used included:
- being led by what is within the software used for cataloguing
- the need to work cross-domain
- the need to be interoperable
- the need to apply very specific subject terms
- the need to follow what the library does
- the importance of an international perspective
- the lack of forethought on how users might use indexes
- the lack of a specialist thesaurus in the subject area the repository represents (e.g. religious orders)
- following the recommendations of the Archives Hub and A2A
* the title of this blog post is a Library of Congress approved subject heading