Methods of Conceptual Scaling

~ Under Construction ~

Orientation

First a word of direction and orientation. What we are discussing here are methods of conceptual scaling. Although these are related to both kinds and types of conceptual scales, they are based primarily upon pragmatic considerations in the practice of conceptual scaling.

Methods

Controlled Vocabularies

These are usually abstract conceptual scales. They are often part of a generic ontology, but not always. Examples include: a collection of genre for movies, an official collection of keywords, the list of manufactured products of a corporation, subject headings, etc. Schema of Dublin Core fit here.Modifier (qualifier) scaling, whose examples include adjectives and adverbs,is a special case.

Direct Scaling

These are always realized conceptual scales (facets), and usually are part of the specification of a specific ontology, either a theory in that ontology or a collection (of attributes) controlled by that ontology. They are manifested as binary ontological relation instances, whose first argument is a particular instance (object) of a category (object type) in the generic ontology being specified, and whose second argument is a value in a controlled vocabulary.

The facet (realized scale) classification relation is the binary ontological relation.

For fixed first argument instance of the relational instance specification, the collection of second argument values defines the state description of the first argument. This method definitely demonstrates why the second step of specification (conceptual scaling) needs to move closer to the first step of specification (ontology construction). Examples include: the genre relation for movies, the keyword relation for press releases, etc. Elements of Dublin Core fit here.

Simple Scaling

This is the classic method of conceptual scaling. Here two ingredients combine to form the state descriptions (rows) of the classification tables: a previously specified domain conceptual scale, either for a datatype or a controlled vocabulary; and a description function from an ontological category to this domain scale.The description function corresponds to a column in a database table.

The facet is the composition of the state space with the concrete domain scale.

This can be the participant part of the bottom-up relational scaling discussed below. Examples include: time period scaled ordinally or interordinally, person's age scale, etc. Equivalence scaling is a special case.

Relational Scaling

A simple example of relational scaling and distributed logics is provided by the File Copying example in the 1997 book Information flow: The logic of distributed systems authored by Jon Barwise and Jerry Seligman. Our development in CKML of this example demonstrates how to do conceptual scaling with respect to relation types, and not just object types.

Complex Scaling

Examples include: use of the residuation operators from relational calculus to defined a universally quantified attribute definition.

Please send questions, comments and suggestions about this page to: Robert E. Kent rekent@eecs.wsu.edu Last modification date: October 1998