Angelina Quan · Fall 2025
Metaphysics
A tempting principle about existence says that, for any property $F$, if $Fx$, then $x$ exists. In ordinary language, this can sound obvious. If something is red, or tall, or a dog, then surely there must be something that has that property. Properties need bearers. So if a sentence truly predicates something of $x$, then it seems that $x$ must exist.
I argue that this principle is too strong. We can meaningfully and sometimes truly ascribe properties to things that do not exist. Fictional and modal discourse both show this. We say that Sherlock Holmes is a detective, that Pegasus is a winged horse, and that a possible child I never have could have had brown hair. These claims do not require their subjects to exist in the actual world. So the principle that $Fx$ always implies that $x$ exists is not a deep truth about predication. It is an artifact of a narrow and overly idealized way of translating ordinary language into logic.
The simplest counterexample comes from fiction. Sherlock Holmes does not exist. There is no actual person living at 221B Baker Street who solved crimes with Watson. Still, it seems true that Sherlock Holmes is a detective. It also seems false that Sherlock Holmes is a basketball player. These claims are not meaningless. They are part of how we understand the fiction. So fictional objects seem able to have properties even though they do not exist.
The same point appears in modal cases. We often talk about merely possible things. For example, I might say that I could have had a younger brother, and that this possible brother would have been younger than me. That does not mean there is an actual person who is my younger brother. It means that there is a possible situation in which such a person exists. Still, the sentence ascribes properties: being my brother, being younger than me, and existing only in a possible scenario rather than the actual world.
These examples suggest that predication does not always require actual existence. Sometimes we predicate properties within a story, a possibility, or a hypothetical situation. The subject does not need to exist in the actual world for the predication to make sense.
A defender of the principle can respond that this confuses ordinary language with strict metaphysics. When we say "Sherlock Holmes is a detective," we are not really saying that there is a non-existent object, Sherlock Holmes, that has the property of being a detective. Instead, we are saying something like: according to the Holmes stories, there is a detective named Sherlock Holmes. On this analysis, the property is not truly being attached to a non-existent object. The sentence is really about the content of a fiction.
This is a strong objection. It explains why fictional discourse does not immediately commit us to strange non-existent objects. It also fits with a more careful logical method. Instead of treating "Sherlock Holmes" as a name for an object, we can treat fictional claims as elliptical claims about stories. Then the principle can be preserved: whenever something really has a property in the strict metaphysical sense, that thing exists.
But this defense comes at a cost. It saves the principle only by heavily rewriting our ordinary and philosophical talk. When people say that Sherlock Holmes is a detective, they are not usually making a complicated claim about what is true according to certain texts. They are directly making a claim about Holmes as a fictional character. The claim is not confused. It is a normal and useful part of our language.
The deeper problem is that the strict metaphysical analysis makes predication too narrow. It treats ordinary claims as misleading unless they can be translated into a form where every property-bearer exists. But that seems backwards. A good theory of predication should explain how our actual predicative practices work. It should not dismiss large parts of them as merely loose or defective.
The same problem appears with modal discourse. Suppose I say, "The child I could have had would have been loved." This is not just a claim about words or stories. It expresses a real modal thought. It says something about how things might have gone. If we insist that predication only applies to actual existents, then we have to paraphrase this away too. We might say: in some possible world, there exists a child who is mine and loved. That paraphrase is useful, but it does not show that the original sentence was illegitimate. It only shows that we can represent it differently.
This is why the principle is not as obvious as it first seemed. It takes one formal setting and treats it as the measure of all meaningful predication. In first-order logic, atomic formulas are often interpreted over a domain of existing objects. So if $Fx$ is true, then $x$ is in the domain. But that is a feature of a formal system. It does not automatically settle the metaphysics of fiction, possibility, or ordinary language.
A defender might say that allowing predication without existence leads to an inflated ontology. If Sherlock Holmes has properties, then are we committed to a realm of non-existent objects? Do we have to say that Pegasus, Hamlet, and every merely possible person are all somehow real? That seems like a serious cost.
This objection is important, but it does not prove the original principle. It only shows that we need a careful account of different kinds of predication. We can say that fictional characters have properties within fictions, and possible objects have properties within possible scenarios, without saying that they exist in the same way ordinary physical objects do. The point is not that Sherlock Holmes exists after all. The point is that meaningful predication has more forms than the simple schema allows.
In conclusion, the principle "for any $F$, if $Fx$, then $x$ exists" is too strong. Fictional and modal discourse show that we can ascribe properties to things that do not exist in the actual world. The defender can preserve the principle by translating these cases into more careful semantic forms, but this makes the principle look less like a deep truth and more like a rule of an idealized notation. A better view is that predication is more flexible. Some predication concerns actual objects, but other predication concerns fictional, possible, or hypothetical objects. Existence matters, but it is not always required for meaningful property-ascription.
|