From the outset, one of the principal aims of the bpi1700 project has been to provide a rich and sophisticated subject index for those whose primary interest is the subject matter depicted in prints. This has entailed constructing a purpose-built thesaurus, hierarchically organised under a series of categories which encapsulate the world and man’s place in it and thus making accessible all subjects appearing within images.

In certain respects, inspiration was derived from such effective bespoke classificatory systems as the Tate’s Insight, or Collage at the Corporation of London, and elements of our classification – including some of the general headings used – reflect these. More important, however, has been ICONCLASS, the iconographic system developed in the Netherlands in the post-war period by Henri van de Waal (1910-72), who aimed ‘to devise a universal system of iconographic classification which could accommodate all representable subjects in Western art’. Since then, ICONCLASS has been widely disseminated in printed and online form and is generally acknowledged to have become the international standard for describing iconography. ICONCLASS breaks up all subjects into the nine main divisions, within which the subjects are subdivided following an alphanumeric format that allows for progressive differentiation and increasingly specific headings. Mindful that, so far as possible, it was not desirable to reinvent the wheel, we have followed ICONCLASS in our classification system wherever this was feasible, adopting its alphanumeric coding for entries even when these have redistributed. (One of the advantages of this is that non-English speaking users can search the bpi1700 database simply by using these codes).

However, though largely adhering to ICONCLASS, significant modifications were made to it to reflect our corpus and its subject matter. In the first place, we have subdivided the category ‘Society, civilisation, culture’, which appears undifferentiated in ICONCLASS, making what are sub-categories in ICONCLASS into full categories in our system and resulting in a slightly longer list of subject headings, reflecting the categories which we thought were likely to interest our users. We have also extended the subject index to include the category of ‘Place’; a further category that may be added at a later date is ‘Events’.

At a more detailed level, we have followed ICONCLASS where appropriate, but we have diverged from it, partly in places where we felt that we needed extra sub-categories to do justice to our material, partly where we felt that categories were misplaced, and partly where ICONCLASS has unnecessary complexities which we thought might be slightly offputting to the user. In particular, ICONCLASS indulges in a degree of subdivision which may seem excessive, sometimes going through eight or nine levels before the required concept is reached. For instance, to reach the womb, ICONCLASS goes through nine levels, whereas we have restricted our hierarchy to four – Human being, the body, anatomy, female sexual organs. In addition, our system was tested as prints were subject indexed, thus complementing the ‘top down’ approach by a ‘bottom up’ one, and this led to a significant amount of modification to the categories that we originally included. Hence what we have tried to do is to produce a system based on ICONCLASS but adapted to the corpus which the bpi1700 database comprises.

A further factor is that successful iconographic indexing requires a controlled vocabulary—a set of standardised keywords and phrases allowing users to locate illustrations matching their search terms even if they had not searched for the exact term given in our subject index, thus mediating between the users and the material. For this, we have used the Library of Congress Subject Headings, perhaps the most comprehensive standard of its kind, which is just over a hundred years old. (The list of subject headings used was first published in 1909-14, but their incorporation into the Library of Congress catalogue went back to 1898, based on an American Library Association list published 3 years earlier.) The system was originally designed as a controlled vocabulary for representing the subject and form of the books and serials in the Library of Congress collection, with the purpose of providing subject access points to bibliographic records. In recent years, it has increasingly been used as a tool in a number of online bibliographic databases, and is now a recognised standard.

Hence we have used this resource to determine preferred subject terms when creating subject headings for this project. For instance, a search for ‘writers’ in the Library of Congress Subject Headings, will reveal that ‘authors’ is the preferred term, and this is what we have used, the use of such ‘controlled terms’ helping not only to ensure consistency of terminology but also to enable compatibility with other projects around the world. On the other hand, in order to direct users from non-preferred terms to preferred terms, we have created a controlled language thesaurus made up of preferred and non-preferred terms, which is built in alongside the subject index by handcrafting ‘see’ terms, particularly taking into account early modern concepts and ideas. For example, a user searching for ‘farm animals’ will automatically be directed to ‘livestock’ in the subject index. The thesaurus also has some element of customisation, including certain headings not recognised by Library of Congress that we need to have in our index, and certain headings where we feel that the preferred terms given by Library of Congress are inappropriate to our early modern material (though these have been kept to a minimum).

As for how images have been subject indexed, we have tried to satisfy the needs of those interested in them by striking a balance between the general and the specific. At one extreme there are general categories like ‘political propaganda’; at the other, specific details like ‘locks and keys’, or ‘sword fights’. But we have also tried to do justice to intermediate levels such as ‘conspiracy’ or ‘faction’. All this means that those who consult the interface can search for any topic that interests them either simply by entering a topic (which is automatically ‘translated’ to the preferred terminology if appropriate), or by browsing – by accessing the subject index with its hierarchies from the general to the specific, and locating the area that interests them -- or by a mixture of the two. The intention is not only to facilitate specific searching, but also, like all good iconographic systems, to encourage users to interrogate adjacent entries through which they are likely to find related material and topics of which they would otherwise have been unaware.

The main headings are as follows:

  • Agriculture, industry and commerce
  • Society
  • Buildings
  • Learning and the arts
  • Nature
  • Places
  • Human being
  • Military and war
  • Politics
  • Religion and belief
  • Bible
  • Mythology
  • Ideas and concepts

Within these, users can follow the subject tree and locate subcategories as appropriate. For example: Society – Family – Marriage – Royal weddings. Alternatively, an alphabetical index of all subject entries is available; there is also a facility to enter a subject keyword.