Value ontologies

Assumptions
Value theory encompasses a range of approaches to understanding how, why, and to what degree humans should value things, whether the thing is a person, idea, object, or anything else. In philosophy the study of values and value systems is called axiology.

The codification of value systems is a challenge for formal ontology (Hartman, 1967). Thus, some departure ontological commitments are required to limit the philosophical concerns behind axiology. Our departure position can be summarized in the following points (see Ott reference below for an introductory view):


 * The ontology does not assume an idealistic paradigm, e.g. there are not built-in sets of ideal and objective values. However, such sets can be embedded in a particular ontology module, separated from the others.
 * The ontology is capable of describing both specific values (individual values, as preferences or interests) and general values.
 * The objective-subjective distinction for values can be avoided by representing a plurality of ontologies. This way, the ontology does not attempt to determine what are objective norms a priori, but represent the different alternatives.
 * The ontology does not include an intrinsic-instrumental differentiation of values, as they are controversial. The decision to describe something as intrinsic (inherent) or instrumental (part of a mean-end relation) is left to the particular value systems codified.

The aim of the ontological schema is being deliberately generic and agnostic to positions concerning the nature of values, this way it can be used to model different possibly incompatible standpoints.

Representation language
The schema is represented in the OWL language, not using currently any features of the OWL 2 version. The following files are provided:
 * values.owl: Basic upper definitions for evaluative concepts, evaluative attributtes and value kinds.
 * intensional-axiology.owl: Basic schema for axiological concepts that are defined by matching the ideal intensional characteristics they should possess. The computation of the levels to which instances match need to be realized either through SWRL rules of through any other computational mechanism external to OWL.
 * oe-axiology.owl: Example definitions for an hypothetical organic agriculture value ontology. It uses an stub ontology with a few organic agriculture definitions oe-ontology-stub.owl

If you use or cite the ontology please cite the following paper:
 * García-Barriocanal, E., Sicilia, M.A. and Sánchez-Alonso, S. (in press). Providing semantic metadata to on-line learning resources on sustainable agriculture and farming: combining values and technical knowledge. Interactive Learning Environments.

The basic schema
Values can either be represented through attributes or concepts (or both). The main values meta-schema provide the basic mechsmisms for both kinds of value representation. The main ontology classes are EvaluativeConcept and EvaluativeAttribute.

subclasses are intended to be defined classes subsumed by non-evaluative concepts. For example, the class of  may be defined as an   that is also a subclass of   and for which some conditions hold (e.g. synthetic fertilizers are not used).

is intended to associate representations of attributes as "sustainable" with instances of any kind. For example, the sentence  relates a concrete fertilization technique with a concrete evaluative attribute.

Both approaches can be combined, especially when there is some definition available. For example, if the ontology provides a way to determine decision rules for "good" techinques, classification can be automated:

In a similar way, rules can be used to assert that a concrete fertilization technique has the sustainable attribute. This example is not intended to be a realistic definition, but to show that evaluative concepts or attributes can be either explicitly declared or be inferred. It is possible also to infer concepts from attribute definitions so that if

Evaluative concept or attribute definitions can be used as markers to search for value-related definitions, and to discriminate them from non-evaluative concepts. It should be noted that the consideration that some attribute or concept is evaluative is left to the modeler, as it might be argued that something being sustainable is objective rather than subjective.

Evaluative concepts can be related to instances of  as a way of structuring clusters of related values. For example, instances x of evaluative concepts related to sustainability may have the.

Ontological plurality
The ontology provides only a high-level schema for defining axiological collections. As it follows formal axiology theories based on attributes of entities, it would typically be importing some domain ontologies. For example, for environmental values it would typically import some agricultural or organic agriculture ontology.

If value collections are kept in differentiated OWL namespaces and modularized in separate .owl files, inconsistent values can be compartimentalized and used separatedly. For example, "good" housing conditions for animals that are incompatible according to different axiological positions can be represented in separate OWL files importing the base schema.