Page 55 - Encyclopedia of Philosophy of Language

ceived clear formulation only in works of post-Fregean philosophy. The issue concerns the existence of, and the indispensability of our reference to, abstract objects. An abstract object (a proposition, say, or a set) is a particular object, but one that possesses nei- ther spatio-temporal characteristics nor causal powers. This is a different issue from the first, because abstract objects are themselves particular individuals: they do not have instances; they do not inhere in substances; and so the problems concerning the nature of universals, and the relation of universals to the particulars that instantiate them have no special per- tinence with respect to them.
See also: Ontological Commitment; Ontology; Uni- versals.
Bibliography
Carr6 M H 1946 Realists and Nominalists. Oxford University Press, London
Field H H 1980 Science without Numbers. A Defense of Nom- inalism.Blackwell, Oxford
Goodman N 1956/4 World of Individuals. The Problem of Universals. University of Notre Dame Press, Notre Dame, IN
Goodman N, Quine W V O 1947 Steps towards a con- structive nominalism. Journal of Symbolic Logic 12:105- 22
Hobbes T 1962 Leviathan Collins/Fontana, London
Loux M J (ed.) 1970 Universals and Particulars: Readings in
Ontology. Anchor Books, New York
The world appears to contain many different things: mountains, people, neutrons, battles, numbers, sets, and so on; and both the variety of these objects and their properties can be investigated. However, lists of these different sorts of things are always controversial: nominalists dispute the claim that there are really abstract objects such as sets and numbers; some pos- itivists refuse to countenance theoretical entities such as neutrons; and other philosophers insist on listing events, such as battles, alongside other objects. Some- body is ontologically committed to objects of a certain kind if such objects must exist for their beliefs about the world to be true. It is often thought that standards of simplicity should oblige people to keep their onto- logical commitments to a minimum; but it is unclear what criteria should be employed in settling whether objects of some kind exist.
1. CriteriaofOntologicalCommitment
It can often be difficult to identify someone's onto- logical commitments. A sentence which appears on the surface to involve reference to objects of a certain kind may be paraphrased in a way that shows this to be misleading. The lost city of the Incas did not exist' appears to involve reference to a (nonexistent) lost city; the paraphrase 'It is not the case that there was a lost city of the Incas' removes this appearance. Many philosophical 'analyses' offer paraphrases which reduce persons' Ontological commitments: analyzing numbers as sets or inscriptions, social institutions as sets of people, and so on. And it is sometimes argued that sentences carry ontological commitments which
are not apparent from the surface. Donald Davidson has argued, controversially, that asserting 'Shem killed Shaun with a knife' refers not only to Shem, Shaun, and the knife but also makes a covert reference to an event of killing (Davidson 1980: ch. 6). So a rule or criterion is needed which, applied to a set of sentences expressing a theory or corpus of beliefs, determines what the ontological commitments of one who accepts those sentences would be. Quine's paper 'On what there is' (1953: ch. 1) is an early attempt to find such a criterion.
Ideally a criterion of ontological commitment would point to a syntactic feature of natural language sentenceswhichalwayssignalsacceptance oftheexis- tence of objects of a certain kind. Use of singular terms will not provide such a criterion: names of nonexistent things are used; definite descriptions do not always have a referring function; and many things exist which cannot be named or described. Quine proposed that ontological commitments are most explicitly signaled by the existential quantifier: I display my commitment to the existence of neutrons by saying There are neu- trons.' Quine has acknowledged the 'triviality' of this view. In fact it can still mislead: there are locutions in ordinary language which appear to have this form but which probably do not carry ontological commit- ments. Thus application of the criterion requires that sentences first be paraphrased into First Order Logic, the Predicate Calculus. A body of sentences carries ontological commitment to objects of a certain kind if its paraphrase into First Order Logic involves quantification over objects ofjust that kind.
Ontological Commitment C. J. Hookway
Ontological Commitment
33