Why classification matters The foundations of bibliographic classification

Some current thinking about classification: • • • it is obsolete it reflects nineteenth century views of the world it is essentially artificial it is a librarian’s device (and therefore to be resisted) it sees everything in terms of hierarchy it has a top-down approach to information • whereas, in reality, • everything is miscellaneous • there is no hierarchy – only links

No hierarchy – only links:

How wrong is this? • the world is not random • there are many observable organized systems in the physical world, with hierarchy and other relationships • taxonomic structures are not invalidated if the basis of the taxonomy changes • even where there is no natural hierarchy, imposing order can only improve findability • various ways of representing subject content support search, query formulation, retrieval, browsing, navigation, awareness of the subject domain, and so on

Pattern and predictability:

Early bibliographic classifications: • • • are often little more than shelf-lists lack structure are full of ‘false’ hierarchies don’t differentiate between kinds of relationships notations may not reflect structure don’t have an analytical approach to content • in short • they lack logical foundations and reflect the absence of information science theory at the time of their creation

Mixed paradigmatic and syntagmatic relationships(SC):

Emergence of classification theory in the 20 th century: • writers such as H. E. Bliss start to formalise classification theory • basic principles for library classification for shelf order and are laid down by him • Otlet establishes classification for documentation, with analytico-synthetic tools designed for retrieval rather than physical organization, and the expression of complex content • Ranganathan introduces the idea of faceted classification, and also develops a substantial corpus of theory • his ideas are taken up by the Classification Research Group who produce a distinctly British version

Modern classification theory: • gives us a proper basis for organizing a subject domain as represented by its constituent concepts • it identifies the difference between concepts, terms, and control devices (such as notations or urls) • (Ranganathan’s idea, verbal, and notational planes) • it shows that, nevertheless these aspects are all interrelated • it provides a common basis for systematic, alphabetic, post-coordinate and pre-coordinate systems • good theory supports the creation of consistent, usable tools

A classification (KOS) consists of: • concepts (vocabulary) * conceptual analysis * categorization (or grouping) * order * relationships • system syntax (grammar) * how to deal with compounds * rules for combination * citation order * ways of maintaining linear order

Modern faceted methodologies build logical systems: • • concepts are organized into facets are organized into arrays (or sub-facets) principles of division are clearly articulated terms in the same facet are all in hierarchical (paradigmatic) relationship terms in different facets are in non-hierarchical (syntagmatic) relationship filing order of facets is determined by principles such as increasing concreteness, general-before-special, and dependency filing order of arrays is more pragmatic, but logically derived order of combination is equally well established methodologically

The structure of the faceted classification: ___________________________________ • it uses ‘simple’ concepts, organized into a logical structure • it uses a standard set of categories to analyse the concepts • it has a standard ‘syntax’ for building compound descriptions • it can be used to create structured classmarks, or • it can provide thesaurus descriptors • it can be used in a search interface

Basic classification structure: [Foods] Hierarchical relationships HKH PO HKH PP HKH PS HKH PY HKH QD HKH QE HKH QF HKH QK Collocation of synonyms HKH QS Facet label (By physical state) Essences Extracts Pastes Array labels (By operation/process used) (By utility, etc. ) Convenience foods Partly prepared foods Instant foods Artificial foods, synthetic foods (By purpose) (By physiological function) Roughage

Complex repeating structure derived from syntax rules: HUQ WH HUQ WMD V HUQ WME HUQ XS Thymus gland (Physiology) (Pathology) (Hyperplasia) Lymphatism, status lymphaticus (Causal agents) (Symptoms) (Treatment) (Neoplasms) Thymomas (Products) Thymus hormones (Molecular structure) Thymopoietins [Compound terms pre-synthesized and added to published schedule] [Examples of potential synthesized compounds]

What are the advantages of a structured KOS? • it is highly logical and predictable • as a result it improves performance • it is a good organizing tool for a physical collection or a linear file • it has a natural affinity with automated systems • the classification data can be held in a database • because of the clear articulation of relationships the data can be output as a thesaurus • a complex structure is easily developed as examples of compounds are added to the system • it is easily exported into other representation formats • it provides a good basis for visualization tools

View-based Systems HIBROWSE search tool: