OntoClean

OntoClean

This following discussion is based on the article [GuWe2004a] of Guarino and Welty: "OntoClean is a methodology for validating the ontological adequacy of taxonomic relationships. ... OntoClean has provided a logical basis for arguing against the most common modeling pitfalls, and arguing for what we have called “clean ontologies”".

Skewed Meanings

Page 2: "Many terms have been borrowed by computer science from mathematics and logic, but unfortunately this borrowing has often resulted in a skewed meaning. In particular, the terms property and class are used in computer science with often drastically different meanings from the original. The use of the term property in RDF is an example of such unfortunate deviation from the usual logical sense". This issue has been addressed in several sections of this work, for example, see Naming Guidelines.

Properties

Page 2: "In this chapter, we shall consider properties as the meanings (or intensions) of expressions like being an apple or being a table, which correspond to unary predicates in first-order logic. Given a particular maximal state of affairs (or possible world), we can associate with each property a class (its extension), which is the set of entities that exhibit that property in that particular world."

This use of properties differs significantly from the use of data properties and object properties used throughout this paper. But this is also due to the Description Logic (DL) notation, because there the unary predicate Person(Tom) or Animal(Bear) is assumed to mean is-a, whereby it cannot be distinguished in DL whether it is the instantiation of an individual instance or a subclass-relation.

Subsumption

Page 2: "The independence of properties from states of affairs gives us the opportunity to make clear the meaning of the term subsumption we shall adopt in this paper. A property p subsumes q if and only if, for every possible state of affairs, all instances of q are also instances of p. On the syntactic side, this corresponds to what is usually held for description logics, P subsumes Q if and only if there is no model of Q ∧ ¬P."

Essence

Page 3: "A property of an entity is essential to that entity if it must be true of it in every possible world, i.e. if it necessarily holds for that entity. For example, the property of having a brain is essential to human beings. Every human must have a brain in every possible world". See also Wittgensteins Definition for internal properties in section LW 4.123.

Page 3: "Obviously there are also non-rigid properties, which can acquire or lose (some of) their instances depending on the state of affairs at hand. ... For example, the property being a student is typically anti-rigid – every in-stance of student can cease to be such in a suitable state of affairs ...".

Rigidity

• Page 3: "A special form of essentiality is rigidity; a property is rigid if it is essential to all its possible instances; an instance of a rigid property cannot stop being an in-stance of that property in a different world. ... On the other hand, the property being a human is typically rigid, every human is necessarily so".
• Page 3: "... anti-rigid properties cannot subsume rigid properties ... For example, the property being a student cannot subsume being a human ..."

Unity

Page 4: "Although very subtle and difficult to explain without experience, identity and unity are perhaps the most important notions we use in our methodology. These two things are often confused with each other; in general, identity refers to the problem of being able to recognize individual entities in the world as being the same (or different), and unity refers to being able to recognize all the parts that form an individual entity.". In [Guar1998] Guarino formally defines the Identity Criterion.

Assigning Metaproperties

• Page 7: "We have chosen, following our previous papers, to use a shorthand notation for indicating metaproperty choices on classes. Rigidity is indicated by R, identity by I, unity by U, and dependence by D. Each letter is preceded by +, - or ~, to indicate the positive, negative, or anti metaproperty, e.g., being rigid (+R), carrying an identity criterion (+I), carrying a common unity criterion (+U); not rigid (-R), not carrying an identity criterion (-I), not carrying a common unity criterion (-U); be-ing anti-rigid (~R) and having anti-unity (~U). We also used (+O) to indicate when a property carries its own identity criterion, as opposed to inheriting one from a more general property". Such an own identity criterion in our approach is given by the used of Unique Data Properties (UDP).
• Page 4: "One of the distinctions proposed by OntoClean is between properties that carry an identity criterion and properties that do not."
• Page 4: The former are labeled with an ad-hoc metaproperty, +I.
• Page 5: "Since criteria of identity are inherited along property sub-sumption hierarchies, a further distinction is made to mark those properties that supply (rather just carrying) their “own” identity criteria, which are not inherited from the subsuming properties. These properties are marked with the label +O (where O stands for “own”)."

Most Abstract Property

Page 8: "Everything is necessarily an entity. Our metaproperties assignment is -I-U+R. This is the most abstract property, indeed it is not necessary to have an explicit predicate for it."

Backbone Constraint Violations

Page 13: "After making the initial decisions regarding metaproperties and arranging the properties in a taxonomy, we are then in a position to verify whether any constraints imposed by the metaproperties are violated in the backbone. These viola-tions have proven to be excellent indicators of misunderstandings and improperly constructed taxonomies. When a violation is encountered, we must reconsider the assigned metaproperties and/or the taxonomic link, and take some corrective action."

Attributions

Page 18: "The final category of properties we consider are attributions. We have one such property in our example, Red, whose instances are intended to be red things. We think that in general it is not useful to represent attributions explicitly in a taxonomy, and that the proper way to model attributions is with a simple attribute, like color, and a value, such as red."

In our approach we would not follow this recommendation in general. In section Color Partitioning we show, that it might be advantageous to model attributions with partition subclass hierarchies.

Page 18: "Although it is not always the case, the cleaned taxonomy has far fewer multiple inheritance links than the original. The main reason for this is that subsumption is often used to represent things other than subsumption, that can be described in language using is a. We may quite naturally say, for example, that an animal is a physical object, however we have shown in this chapter that this kind of linguistic use of is a is not logically consistent with the sub-sumption relationship. This results in many subsumption relationships being removed after analysis"