Between symbol and language-in-use

Indexing is often designed with the intent of dimensional reduction, that is, of generating standardised and uniform descriptive metadata. This could be characterised as a process of decontextualisation. Formal knowledge representation systems typically have the aim of encapsulating granular pieces of information in a reusable manner. The result is a set of information elements with minimal links to external information sources. Plain-text tags, by comparison, have the aim of describing an object, within or outside a reductively described context. The result is a set of views that are contextualised to author, time, location, task or community. This paper discusses the relationship between symbol, contextual relation and language-in-use.