Can there be a criterion of ontological commitment? Can there be a formal test of what a person is ontologically committed to? What a person is committed to is a matter of what he believes or assumes or presupposes or is prepared to act on—on his attitudes. So the question is whether there is a linguistic litmus test for an attitude of commitment. Can we read a person’s ontology off his verbal productions? Can I figure out my ontological commitments by inspecting my use of language?
The first thing to observe is that the question is not restricted to matters of existence. As the term is commonly used “ontological commitment” is taken to refer to what a person takes to exist, so that it is interchangeable with “existential commitment”. That is certainly one form of commitment—what a person believes to exist—but it is not the only form. Consider “chromatic commitment”: what colors you believe things have (whether they exist or not). You may believe that things are colored and you may believe specific color claims—these are your chromatic ontological commitments. Ontology concerns what is so, and color is a matter of what is so. Roses are red and violets are blue—and Santa Klaus has a white beard and a red cloak (whether he exists or not). I might believe that colors are unreal and that nothing has them; in that case I am not ontologically committed with respect to color, though I might well believe in the existence of the things commonly said to be colored. Ontological commitment can concern any fact or putative fact: do you believe in that fact or not? Do you believe in moral facts, divine facts, facts about unobservable entities, psychological facts, and so on? Existence is just one kind of ontological commitment: we might say that it concerns one type of property, viz. the property of existence. Does anything have the property of existing? Which things do? Does anything have the property of being colored? Which things do? And so for any property you care to mention. A criterion for existential commitment might be a willingness to affirm “Such-and-such exists”, and a criterion for chromatic commitment might be a willingness to affirm “Such-and-such is red” (and similarly for other kinds of fact). It is artificial to single out existence from other sorts of ontological commitment: it is just one kind of factual commitment. The proper contrast here is with “epistemological commitment”: what we are committed to in the way of knowledge. What is it that we think we know? Do we think there is any knowledge, and if so what is known? We can be committed on questions of being (fact, reality) and we can be committed on questions of knowledge; what we are committed to existentially is just a special case of a more general question.
The question of providing a criterion of ontological commitment is thus broader than that of providing a criterion of existential commitment. Quine announced, “To be is to be the value of a variable”; he has been paraphrased thus, “What you say there is, you say there is”. That is, you are committed to whatever your sentences mean: if you affirm a sentence that can be true only if certain things exist, then you committed to the existence of those things. For example, you can’t say, “There are numbers” and then turn round and deny there are numbers: you must be taken at your word. But it is the same with all forms of ontological commitment: if you say, “Roses are red” you can’t turn round and deny that roses are red (same for “good”, “solid”, “conscious”, and so on). To be committed to red things is to describe things as red. You are committed to such facts as your sayings require for their truth. The criterion of commitment is saying. You can’t disavow what you affirm: you can’t say it and then try to take it back. You can’t say it in practice but then disavow it theoretically. You can’t have your ontological cake and eat it. You can’t weasel out of your statements.
That sounds all very reasonable (indeed trivial—what was the fuss all about?), but actually it runs into difficulties as a formal test of ontological commitment. The idea was to provide a public formal test of ontological commitment, eschewing the vagaries of what a person internally believes. We might think of it as a behavioral criterion for a mental phenomenon: what a person is committed to (believes to be) is what he affirms in his public utterances. A person believes in unicorns if she affirms, “There are unicorns” or “Unicorns exist”. I determine what I believe in by examining what I say, and I might be surprised at what turns up (I may find that I accept, say, an ontology of events or possible worlds). Thus the criterion is formal and public: it invokes facts of language and it is interpersonally accessible. No need to delve into the inner recesses of a person’s mind.
But the proposal is obviously problematic. It hardly provides a necessary condition, since you can keep silent about what you believe or may not have language at all; and it is not sufficient, since speech is not always sincere assertion. It is possible to say something and not believe what one says, as in play-acting or elocution practice. Even in assertion you may not be committed to what you assert in the sense that you believe what you say. A liar can’t use his assertions to figure out his ontological commitments. The assertion must be sincere, i.e. you must believe what you assert. But that is what we were seeking a criterion for–belief. Speech is never a sure guide to belief, so we can’t formulate a test of ontological commitment from facts about speech. My ontological commitments can be read off my sincere assertions—if I sincerely assert, “Snow is white”, then I am committed to snow being white—but the commitment comes from the belief not the assertion. No act of speech (or writing) can add up to belief, so there cannot be a formal linguistic criterion of ontological commitment. In order to find out what I am committed to you have to find out what I believe; what I say isn’t going to get you there. It may be true that what I say there is I say there is, but it doesn’t follow that that is what I believe there is. The most that can be claimed is that we have criterion for the ontological commitments of what someone says—a speech act is “committed” to what is required for its truth—but this is a far cry from the ontological commitments of a person. What I believe is not the same thing as what I say, since I may not give voice to my beliefs and, if I do, I may not mean what I say. My ontological commitments are fixed by my beliefs—but that is a trivial tautology not an illuminating criterion.
There is also the case of a speaker actively rejecting the ontology of his sincere assertions. Suppose Meinong is right about definite descriptions—they really do denote non-existent subsistent objects. Then whenever Russell or anyone else makes a statement involving a definite description his speech act is committed to such objects: he accepts the truth of the uttered sentence and its truth requires non-existent objects (in the empty case). But Russell himself vehemently rejects such Meinongian objects—he doesn’t believe in them no matter what his utterances may entail. In this case ontological beliefs cannot be read off sincere assertion plus correct semantic analysis. The same is true for any type of statement: the speaker may reject what his sentences semantically entail. He is not committed to what his sentences are committed to, i.e. what is required for their truth. He may regard those sentences as logically defective and explicitly reject their entailments; they can’t force beliefs upon him. There is always a logical gap between language and belief, so cannot derive a criterion of ontological commitment from features of language. Perhaps non-linguistic action could supply such a criterion (think of animal ontological commitment), but what we say can never constitute what we believe. What I say there is may never be what I believe there is, and similarly for every other type of fact. Ontological commitment is a matter of private belief not public utterance.