Meta‐ and object‐language for information retrieval research: proposal for a distinction

Abstract

The distinction between a meta‐ and an object‐language has become increasingly familiar in information science, through the diffusion of the concept of metadata. A significant antecedent to this distinction can be found in the development of formal logic. This paper proposes an analogous distinction for information retrieval research, between the metalanguage of discourse about information retrieval systems and the object‐language of transformations within systems. In formal logic, acceptance of a meta:object distinction has had a clarifying and simplifying effect. An understanding of potential object‐language transformations as the writing, erasure, and substitution of symbols has also been developed. The existing metalanguage of information retrieval research has displayed a founding assumption (the value of delivering all, and possibly only all, the records relevant to a given query), some central concepts, entities for evaluative purposes, and derived measures. An alternative founding principle of enhanced informed choice is endorsed. The emerging view of operations within information retrieval systems such as transformation, sorting, and partitioning is strongly analogous to the more fully established account of possible object‐language transformations in formal logic. Analytical clarity has been obtained and economy in research effort is made possible.