April 07, 2026; most recent update: April 28, 2026
Table of Contents
1. Introduction
2. Refutations & Clarification
a. The Concept Immune to "Existence Is Not a Predicate"
b. Two Aspects of Existence -- A Likely Factor in the Debate
c. Other Examples Showing Existence is a Predicate
Immanuel Kant's impact on Western philosophy can't be overstated.
One of his claims that has had a lasting influence is in the areas of logic and metaphysics is that existence is not a predicate. It's a very counterintuitive claim and so, as with many of Kant's ideas, has generated a lot of controversy. After all, if there's a chair outside on the patio, it's natural to say that it has the property of existence and that distinguishes it from being just an imaginary chair.
But Kant's idea has some merit. For example, the truth that a triangle is an object with three sides is true regardless of whether or not there are any objects in the universe with those properties. Therefore the existence of any triangles is irrelevant to the properties the concept has and, in turn, the existence of any members isn't because of their triangularity.
The key assumption behind this idea is simple: the properties of a concept are one thing; whether the concept has any real members, another.
It's a crucial difference since the properties of a concept are necessary and eternal, whereas the existence of actual examples is an arbitrary fact that may or may not be the case. Thus no concept -- not even concepts pertaining to single, existing things (the Eiffel Tower, etc) -- has members that necessarily exist. For instance, it might seem that for there to be a concept of the Eiffel Tower, its member must exist: the concept necessarily refers to an actual thing. But people can still talk of the Eiffel Tower in the future even if it doesn't exist. So, a concept's members necessarily have the properties of the concept, but existence is never one of those properties. Therefore, existence isn't a predicate.
Kant regarded existence as a modality rather than a predicate. He recognized three sets of modality: existence/non-existence, possibility/impossibility, and necessity/contingency.
Kant was responding to ontological arguments. Such arguments are essentially that the concept of God entails some superior quality (e.g. the greatest being or a perfect being), lacking existence would be a lack of that quality (being the greatest or being perfect, etc), and therefore the concept's member necessarily exists.
While Kant's claim is true of some concepts, it isn't true of all of them.
There's one counterexample that is so obvious that it can create hesitation in giving it: if it's a legitimate exception, how could it have been missed by so many philosophers for so long? Nevertheless, there doesn't seem to be an adequate response from Kant's position.
What about the concept of existent? The only thing that its members necessarily have in common is the property of existence. To still maintain that this "property" is just a modality and not a real predicate would be to claim that the concept of existent doesn't have any property. For it is that "property" alone that unites its members. And to claim that the concept lacks properties would entail something even more amazing. It would mean we have a propertyless concept that somehow we are still able to distinguish from other concepts.
Clearly those are unacceptable implications. EXISTENT necessarily gives the predicate of existence to its members.
Still, while members of EXISTENT necessarily have the predicate of existence, they don't exist unconditionally: their existence isn't necessary in the sense of them existing "in all possible worlds," to use a phrase from Leibniz. Thus existence has two aspects: as a predicate and as a modality. So, Kant wasn't totally wrong.
It's likely that these two aspects of existence have caused many people to talk past each other during dialogue over whether existence is a predicate, leading to confusion and disagreement, thus further fueling the debate.
The counterexamples certainly don't end with EXISTENT.
[Necessary Locations]
---------------------
A second type of counterexample involves necessary-location concepts. For example, whatever area of geography you choose, there is necessarily a point that is furthest north in that area, furthest east, etc. Such points necessarily exist in that area, and therefore exist so long as the area does too. Because they necessarily exist in that area, they also have the property of existence.
This kind of concept isn't restricted to planets. As long as there is at least one object with spatial dimensions, there will necessarily be a location furthest in front of it, behind it, to the left of it, to the right of it, above it, and below it. Even if space is limitless, such locations can still be established by conceiving of a limited, hypothetical area with the spatial object in the center of it. Which way was which -- above, below, right, etc -- would be in the eye of the beholder, but that still wouldn't stop such locations from necessarily existing in the given area.
Another way of seeing whether or not existence is a predicate is to ask if
its inclusion or exclusion in a definition would affect our ability to identify the concept's members. If our ability is affected, whether positively or negatively, it shows that existence is an attribute.
Several kinds of concepts have this effect on us when we add or remove the idea of existence from their definitions.
[Things That Once Existed]
--------------------------
Our ability to identify members is affected when it comes to concepts involving things that once existed but don't anymore.
Take the example of thylacine, otherwise known as the Tasmanian tiger or Tasmanian wolf. A concise definition would be something like: an extinct dog-like marsupial with stripes, native to Tasmania, Australia, and New Guinea. Is the word "extinct" informative? Without it, you might not recognize the animal the definition is meant to reference. A primitive, unsophisticated AI could easily get confused. Suppose we leave out "thylacine" and "extinct" and simply type into the search bar "a dog-like marsupial with stripes, native to Tasmania, Australia, and New Guinea." Our definition would reasonably imply that such an animal exists, and we can imagine an unsophisticated AI responding "no results found." But if we included the word "extinct," it's a good bet that even our second-rate AI would be able to give us what we're looking for. Or, at least the AI's chances of helping us would be much greater. We can say the same for humans given the same task.
Likewise, we can't define the denarius (the currency of the Roman Empire for most of its existence) as a coin currency that is used in parts of West Europe, North Africa, and West Asia. The present tense makes the definition a lie.
So, when it comes to animals that are extinct or other things that have exited the world, not including "extinct" or some synonymous word or phrase makes for a worse definition. It adds some confusion into what things we're looking for since it implies existence.
[Make-Believe Things]
---------------------
Good definitions of make-believe things also depend on mentioning that they don't exist. Below are several examples, each with a worse and better definition.
Unicorn
1. A horse with a single horn.
2. A fantastical horse with a single horn.
Pegasus
1. A horse with wings.
2. A mythological horse with wings.
Leprechaun
1. A diminutive sprite, usually clad in green.
2. A diminutive sprite in Irish folklore, usually clad in green.
We can easily see why the first definition of each set is inferior to the second. The first definitions are misleading at best. Suppose a child were to ask about any of those concepts. The first definition in each set would indicate to the child that such creatures were real. But the members exist only in the realm of imagination. The child's search for unicorns and pegasuses would come up empty when looking through literature on horses.
[Things That Exist]
------------------
But the same general point applies to concepts referring to currently existing things. We have to get the members' status of existence right in the definition. Here's a few examples:
Lion -- an imaginary, large, tawny cat that hunts in groups.
Tiger -- a large, agile, make-believe cat with stripes.
Wolverine -- an extinct, aggressive, badger-like animal capable of climbing trees.
Again, if existence isn't a predicate then it shouldn't matter whether we define a concept as having members that exist or not. Therefore the definitions above should be fine. But they're not, because they prevent us from locating the concepts' members.
It's clear that all of these exceptions to "existence isn't a predicate" cover a very broad range of things. And yet this isn't necessarily an exhaustive list. Our ability to identify members might be affected when we add or remove existence to the definitions of other kinds of concepts as well.
