The term ontological commitment is used as a general term in both philosophy and in information systems to refer to the essential elements of an ontology. An ontological commitment in describing ontological comparisons is taken to refer to a subset of elements of an ontology that it shares with all other ontologies based upon the same theory or conceptualization (see next section for more detail). Quine proposed that, given some theory, its ontological commitment could be found by what might be called a translation via techniques of symbolic logic and a search through this translation for statements involving there exists at least one ‘such-and-such’. Such statements are called quantifier expressions and the formulation ‘there exists’ in symbolic logic is represented by the 'turned E' or ∃. A list of the ‘such-and-such’ can then be examined to determine subsets that can serve as minimal sets in terms of which the others can be defined, and any such minimal set is an ontological commitment of the theory. This approach appears to involve only a list of ‘such-and-such’, but of course finding a minimal set of ‘such-and-such’ also involves at least some of the relations specified to hold between them.
This view of ontological commitment leaves considerations of what really is to ontology and focuses upon what amounts to a linguistic analysis. It is more an issue of semantic categories.
In information science, a conceptualization is an abstract simplified view of some selected part of the world, containing the objects, concepts, and other entities that are presumed of interest for some particular purpose and the relationships between them.
An ontology provides an explicit specification of a conceptualization, and it may occur that a conceptualization can be realized by several distinct ontologies. "An ontology is language-dependent, while a conceptualization is language-independent." Guarino elaborates on what he means by 'language independent' to say that a conceptualization is always the same, "independently of the language used to describe it". Not all workers in knowledge engineering use the term ‘conceptualization’, but instead refer to the conceptualization itself as an overarching ontology.
Purpose and implementation
As a higher level abstraction, a conceptualization facilitates the discussion and comparison of its various ontologies, facilitating knowledge sharing and reuse. Each ontology based upon the same overarching conceptualization maps the conceptualization into specific elements and their relationships.
The question then arises as to how to describe the 'conceptualization' in terms that can encompass multiple ontologies. This issue has been called the 'Tower of Babel' problem, that is, how can persons used to one ontology talk with others using a different ontology? This problem is easily grasped, but a general resolution is not at hand. It can be a 'bottom-up' or a 'top-down' approach, or something in between.
However, in more artificial situations, such as information systems, the idea of a 'conceptualization' and the 'ontological commitment' of various ontologies that realize the 'conceptualization' is possible. The formation of a conceptualization and its ontologies involves these steps:
- specification of the conceptualization
- ontology concepts: every definition involves the definitions of other terms
- relationships between the concepts: this step maps conceptual relationships onto the ontology structure
- groups of concepts: this step may lead to the creation of sub-ontologies
- formal description of ontology commitments, for example, to make them computer readable
An example of the problems encountered in comparing ontologies is found in translation between human languages. Ostensibly, as all humans live in the same world and have the same physical senses with which to see the world, one might expect to correlate human activity with language and thereby make rules for translation. However, that view is utopian because humans act upon cultural interpretation of their surroundings, and relating two cultures is an entirely different matter than understanding what term in both represents a 'rabbit'. Some suggest that humans think in 'mentalese', but so far we don't have access to this level of conceptualization.
However, in more artificial situations, such as information systems, the idea of a 'conceptualization' and 'ontological commitment' to various ontologies that realize the 'conceptualization' is possible. A trivial example of moving conception into a language leading to a variety of ontologies is the expression of a process in pseudocode (a strictly structured form of ordinary language) leading to implementation in several different formal computer languages like Lisp or Fortran. The pseudocode makes it easier to understand the instructions and compare implementations, but the formal languages make possible the compilation of the ideas as computer instructions.
Another example is mathematics, where a very general formulation (the analog of a conceptualization) is illustrated with 'applications' that are more specialized examples. For instance, aspects of a function space can be illustrated using a vector space or a topological space that introduce interpretations of the 'elements' of the conceptualization and additional relationships between them but preserve the connections required in the function space.
- Nicola Guarino (1998). “Formal Ontology in Information Systems”, Nicola Guarino, ed: Formal Ontology in Information Systems (Proceedings of FOIS '98, Trento, Italy). IOS Press, pp. 3 ff. ISBN 978-90-5199-399-8.
- Frank van Harmelen. Ontology mapping: a way out of the medical tower of babel.
Roger F. Gibson (1999). “Ontological commitment”, Robert Audi, ed: The Cambridge Dictionary of Philosophy, Paperback 2nd, p. 631. ISBN 0521637228. A shortened version of that definition is as follows:
- The ontological commitments of a theory are those things which occur in all the ontologies of that theory. To explain further, the ontology of a theory consists of the objects the theory makes use of. A dependence of a theory upon an object is indicated if the theory fails when the object is omitted. However, the ontology of a theory is not necessarily unique. A theory is ontologically committed to an object only if that object occurs in all the ontologies of that theory. A theory also can be ontologically committed to a class of objects if that class is populated (not necessarily by the same objects) in all its ontologies. [italics added]
Luigi Ceccaroni, Myriam Ribiere (2002). "Modeling utility ontologies in agentcities with a collaborative approach". Proceedings of the workshop AAMAS.
A quotation follows:
- “Researchers...come from different areas of study and have different perspectives on modeling, but significantly they pledged to adopt the same ontological commitment. That is, they agree to adopt common, predefined ontologies...to express general categories, even if they do not completely agree on the modeling behind the ontological representations. Where ontological commitment is lacking, it is difficult to converse clearly about a domain and to benefit from knowledge representations developed by others... Ontological commitment is thus an integral aspect of ontological engineering.” [italics added]
- For a discussion and a critique see Peter van Inwagen (2008). “Chapter 6: Quine's 1946 lecture on nominalism”, Dean Zimmerman, ed: Oxford Studies in Metaphysics: Volume 4. Oxford University Press, pp. 125 ff. ISBN 0191562319.
- Dag Westerståhl (Apr 19, 2011). Edward N. Zalta, ed.):Generalized Quantifiers. The Stanford Encyclopedia of Philosophy (Summer 2011 Edition).
- Amie Thomasson. Edward N. Zalta, ed:Categories. The Stanford Encyclopedia of Philosophy (Spring 2013 Edition).
- Gruber, Thomas R. (June 1993). "A translation approach to portable ontology specifications". Knowledge Acquisition 5 (2): 199–220.
- Barry Smith (2003). “Chapter 11: Ontology”, Luciano Floridi, ed.: Blackwell Guide to the Philosophy of Computing and Information. Blackwell, pp. 155-166. ISBN 0631229183.
- For example, see Luigi Ceccaroni, Myriam Ribiere (2002). "Modeling utility ontologies in agentcities with a collaborative approach". Proceedings of the workshop AAMAS.
- In information science, one approach to finding a conceptualization (or avoiding it and using an automated comparison) is called 'ontology alignment' or 'ontology matching'. See for example, Jérôme. Euzenat, Pavel Shvaiko (2007). Ontology Matching. Springer. ISBN 3540496122.
- Nicola Guarino, Massimiliano Carrara, Pierdaniele Giaretta (1994). "Formalizing ontological commitments". AAAI 94: pp. 560-567.
- Maja Hadzic, Pornpit Wongthongtham, Elizabeth Chang, Tharam Dillon (2009). “Chapter 7: Design methodology for integrated systems - Part I (Ontology design)”, Ontology-Based Multi-Agent Systems. Springer, 111 ff. ISBN 364201903X.
- Willard v. O. Quine (2013). Word and Object, New. MIT Press. ISBN 9780262518314. Quine raised the issue of translation and 'holophrastic' indeterminacy of translation in a series of books and papers.
- Crispin Wright (1999). “Chapter 16: The indeterminacy of translation”, Bob Hale, Crispin Wright, eds: A Companion to the Philosophy of Language. Wiley-Blackwell, p. 397. ISBN 0631213260. "Quine's contention that translation is indeterminate has been among the most widely discussed and controversial theses in modern analytical philosophy."
- Murat Aydede (September 17, 2010). Edward N. Zalta, ed:The language of thought hypothesis. The Stanford Encyclopedia of Philosophy (Fall 2010 Edition).