by Jesse Card

Created on: January 06, 2010

If you have ever used language translation software and come up with a translation that does not come close to what you wanted to say in the target language, you have encountered one of the basic problems that computer science is trying to address with the concept and creation of linguistic ontologies.  There are many items which need be defined and discussed to fully explain everything involved, but this article will only briefly address the necessary components.  In the most basic terms, a linguistic ontology is a system which categorically defines all aspects of language with a goal toward creating interoperable databases of the world’s languages.

Ontology as a philosophical concept is branch of metaphysics which is concerned with the study of being.  There are many concepts which are argued within ontology - such as whether existence or expression of existence is the base of being, how objects interact, and the manner in which an object is perceived – which have been debated since the earliest times in philosophy.   A subset of ontology is the attempt to categorize all things in existence and the expressions of existence in a taxonomical manner as scientists do with living objects.  The obvious problems are that 1) the definition of an object or its expression is tied to its observation, which can vary for different people at different times and places (their relationship to the object) and 2) everyone has their own opinion as to what constitutes a category, or even if categories exist.

Computer science encountered similar problems in developing the rules systems use to compute data.  A computer’s basic role is to model existence.  It takes data which a user inputs, translates it by the rules and definitions established by its designers, and presents that data in whatever manner it is instructed to use (i.e. a document, picture, video, audio clip, and every other function it may perform).  The problems arise when different hardware and software use different rules and definitions in processing data (such as creating a document in WordPerfect instead of Microsoft Word or using a Mac instead of a PC). 

This problem is called interoperability, and it exists in all fields of computing and at all levels.  The difference between ontology to philosophers and ontology in computer science is that there is nothing to force a philosopher to make a decision about categories and definitions, but

• 1
• 2
• Next Page >>