Do You Exist..? What Exactly Is Existence?
When you say “I exist,” what exactly are you claiming? That there’s a physical body at a particular location? That there’s a continuous stream of consciousness? But presumably you'd be you with a prosthetic arm, and that stream gets interrupted every night when you sleep.
We’re confident about our own existence, yet we can’t articulate exactly what it is we’re so confident about other than unspecific references to “patterns”.
Today, let’s analyze existence.
To make this more palatable and less psychiatric hospital-inducing, let’s distance this analysis from the first person Dark Night and talk about the Dark Knight.
“Batman doesn’t exist.”
Firstly… sorry to break it to you.
Secondly, this is such a elementary sentence with penetrating and wide implications, not just for philosophy but for your existence. That Batman sentence is the philosophical equivalent of dividing by zero. It crashes the whole system.
Why? After all, clearly Batman’s just pixels and ink, a character in a story, a fictional vigilante who fights crime in a made-up city, etc. No controversy here.
Well, let’s think about it: If Batman doesn’t exist, then who exactly are we talking about? What is this “Batman” that we’re assuredly declaring nonexistent? It’s like saying “This sentence is false” or “I’m not thinking about how devastatingly handsome Curt is.” (Too late, you just thought about my chiseled jawline.)
The Paradox That Literally Ruins Everything
Here’s the problem in logical form. When you utter what philosophers call a negative singular existential (a sentence denying that some particular individual exists), you appear to be:
Using a singular term (a word that purports to pick out one specific individual, like “Batman” or “Santa”) to refer to something
Or predicating the nonexistence of the referent of that term (predicating just means attributing a property to something, for instance, “the apple is red” predicates redness of the apple. A referent is what a term points to, e.g., what “Batman” picks out)
But just a moment… how can a singular term (successfully) refer if there’s nothing to refer to? That’s like successfully pointing at nothing. If you point at nothing, are you pointing at something? What does this even mean? Let’s consider these negative singular existentials:
“Santa Claus doesn’t exist”
“Pegasus doesn’t exist”
“Your Canadian girlfriend doesn’t exist”
“The proof that I’m the most sagacious philosopher alive doesn’t exist” (clearly false)
If these sentences are meaningful and true, then language has a superpower: the ability to talk about nothing as if it were something. That’s either considerably impressive or considerably broken.
Curt’s note: There is the alternate view that language doesn’t refer to anything, but that’s not the route we’re taking here. If you’d like to explore that, then watch the below with
Russell’s Solution: Existence is a Second-Order Property
Bertrand Russell had a solution so dexterous it dominated philosophy for 60 years. Russell suggested that existence isn’t a property of individuals at all. Existence is a second-order property.
What’s a second-order property? Think of it this way:
First-order properties are those like being red, being tall, being happy (so, properties that individuals have)
Second-order properties are those like being instantiated, being widespread, being rare (so, properties that properties have)
You can think of second order properties as meaning “property of properties.” It’s not precisely this, but this will suffice for the sake of the uninvolved task of understanding mere existence.
Curt’s note: Instantiated means “has at least one instance.” For instance, the property “being a ballerina” is instantiated (because ballerinas exist), but “being a unicorn” isn’t instantiated (because unicorns don’t).
According to Russell, when you say “Dragons exist,” you’re decidedly not saying that certain individuals have the property of existence, even though you think you are. Instead, you’re saying that the property being a dragon has the second-order property of being instantiated. In logical notation, we (incorrectly) believe we mean:
(there exists an x such that x is a dragon—the ∃ is the existential quantifier, meaning “there exists”).
Okay, now what about names like “Batman”? Russell’s claim is that they’re not genuine singular terms at all. They’re disguised definite descriptions; that is, phrases of the form “the so-and-so” that pick out a unique individual by description. “Batman” is shorthand for something like “the caped crusader of Gotham.”
This means “Batman doesn’t exist” has a completely different logical form (the underlying logical structure) than its grammatical form (how it appears in English) suggests. To Russel, “Batman doesn’t exist” is just shorthand for:
Translation: “It’s not the case that there exists exactly one caped crusader.” (The formula uses the universal quantifier ∀ meaning “for all,” and → for “implies.”)
The magic of Russell is that he’s just eliminated singular reference to Batman. You’re just denying that a certain property (i.e., being the unique caped crusader) is instantiated. No reference to the nonexistent required.
Kripke Drops the Modal Bomb
For decades, almost everyone thought Russell had solved it. Then in the 1970s, Saul Kripke published Naming and Necessity and demonstrated that Russell’s “descriptivist theory” of names was subtly and fatally flawed.
Curt’s note: “Subtly and Fatally Flawed” is also the alternate name of your favorite interpretation of quantum mechanics.
Kripke pointed out that names and descriptions behave differently in modal contexts (contexts involving possibility and necessity, and sentences with “may,” “could,” “must,” “necessarily”). Consider:
“Peter Parker may not have been bitten by a radioactive spider” — TRUE
“The guy bitten by a radioactive spider may not have been bitten by a radioactive spider” — FALSE (that’s contradictory)
If “Peter Parker” just meant “the guy bitten by a radioactive spider,” these sentences should have the same truth value. They don’t. Therefore, names aren’t disguised descriptions.
Instead, it seems like names are more like rigid designators, picking out the same individual in every possible world where that individual exists. Descriptions ordinarily aren’t rigid as they can pick out different individuals in different possible worlds.
Think about it like this. If “Shakespeare” is synonymous with “the author of Romeo and Juliet,” then anyone competent with English would recognize that “Shakespeare wrote Romeo and Juliet” is an analytic truth (true just by the meanings of the words, exactly like “bachelors are unmarried males”). However, this isn’t exactly the case since it takes empirical investigation to discover who wrote what. Indeed, some people think Francis Bacon wrote Shakespeare’s plays.
Another example is that I can successfully refer to Feynman while believing he was a famous bongo player who invented the hula hoop. I’m wrong about everything except his name, but I’m still talking about Feynman (the physicist)! This is why names aren’t equivalent to the descriptions speakers associate with them.
Kripke’s conclusion: Names are directly referential (they refer without going through this discursive side path of “descriptive content”). In other words, names don’t have descriptive meaning. Names directly point straight at their bearers like a philosophical laser pointer.
Okay, but now we’re back to our original problem, except more abjectly, since if “Batman” is a directly referential term, well… what happens when there’s nothing to refer to?
It’s like a laser pointer aimed at empty space, except we somehow all know exactly where it’s pointing.
The Austrian Solution: Nonexistent Objects
Enter Alexius Meinong, who proposed something so awry that Russell spent years trying to refute it: Perhaps the domain of quantification (the collection of everything we can talk about and quantify over (basically, what “everything” ranges over)) includes things / objects that don’t exist.
According to Meinong, existence is a genuine property of individuals, but not everything has it. This is the more intuitive answer that you probably already have in the back of your mind.
While everything can be an object that we can talk about and say statements about, only some of these objects also have the property of existence. So being an object doesn’t require existing. Seems obvious.
The comprehension principle for objects (the principle that determines what objects there are) says: For any set of properties, there’s an object having exactly those properties. Some objects exist (e.g., the USB C cord on my desk), others don’t (e.g., Batman).
On this view, “Batman doesn’t exist” expresses a perfectly coherent proposition. It says: That object designated by “Batman” lacks the property of existence. Batman is in the domain of quantification since you can refer to him and predicate properties of him, yet he doesn’t exist. He’s an object without existence.
Curt’s note: Non-existence is clearly not the same as non-importance. However, some do infer something’s non-importance from the non-existence of something else, e.g., academics do this exactly for those who don’t have h-indexes.
Basically, to Meinong, everything you can meaningfully talk about is an object, but only some objects exist. Existence becomes a discriminating property like being red or being tall. Some objects have it, some don’t. Simple.
So what’s the problem?
Meinong’s Terrible, Horrible, No Good, Super Bad Day
Russell pointed out some awkward consequences of letting nonexistent objects into your ontology.
Consider the “round square.” By Meinong’s naive comprehension principle, there’s an object that’s both round and square. But that violates the principle of noncontradiction (the logical law that nothing can have incompatible properties). So Meinong’s theory entails contradictions!
Furthermore, what about the “tall man” (an object whose only properties are being tall and being a man)? Well, is he bearded or clean-shaven? Neither. As stated, he’s “incomplete” with respect to facial hair. But that would violate the principle of excluded middle (everything must be either P or not-P). Every object should either have or lack any given property, at least according to Russell.
Curt’s note: Keep in mind there are alternative views to this such as intuitionistic logic and dialetheism (allowing contradictions). I’ve spoken to Graham Priest, the primary developer of this latter sort of logic, below.
The last problem that Russell had with the Meinongian view is that the “existent golden mountain,” characterized as golden, mountain-shaped, and existent. By the naive comprehension principle, there’s an object with exactly these properties. But then it exists! We’ve just thought something into existence.
Modern Meinongianism: Now With Safety Features
Here’s where it starts to get refined. Most modern Meinongians have developed sophisticated versions of Meinongianism that avoid these problems. For instance:
Nuclear/Extranuclear Strategy (Terence Parsons): This divides properties into two types. Nuclear properties, which are ordinary properties that characterize what objects are like (being red, being tall, being dashingly good-looking, etc.). And extranuclear properties, which are higher-order properties about an object’s metaphysical or logical status (existing, being possible, etc.)
The comprehension principle only applies to nuclear properties. Since existence is extranuclear, you can’t define something into existence. The round square has the nuclear properties of roundness and squareness, but it has the extranuclear property of being impossible.
Then there’s the refinement of Edward Zalta’s dual copula strategy that objects can have properties in two manners:
Encoding: Having a property as part of one’s nature or definition (like how a character in a novel “has” properties)
Exemplifying: Actually instantiating a property in reality (like how you actually have properties)
Holmes encodes being a detective but doesn’t “exemplify” it. Instead, Sherlock exemplifies being fictional. The round square encodes roundness and squareness without exemplifying either. So, no contradiction.
These solutions work, but now we need spreadsheets to track which properties are which type and which mode of predication is in play. Meaning that philosophy has become accounting, except without job prospects.
Of course, there’s the more recent approach of accepting that all genuine singular terms refer and that everything exists, but insist that apparently empty names like “Batman” refer to abstract artifacts created by humans. This means that “Batman” refers to an abstract object created by DC Comics. This abstract object exists in the same way numbers and stories exist, so when people say “Batman doesn’t exist,” they’re expressing the true proposition that Batman isn’t “concrete.”
This respects direct reference (names refer without description) as well as classical logic (the standard logic where every meaningful sentence is true or false) while avoiding Meinongian nonexistents. The cost, however, is that it’s revisionary about ordinary usage!
When someone says “Santa doesn’t exist,” they’re saying Santa doesn’t exist in any manner whatsoever. They’re not usually making some subtle distinction about abstract versus concrete existence.
Descriptivism: The Zombie Theory
Some philosophers refuse to let descriptivism die. They’ve developed versions that survive Kripke’s attacks. The three ones that I only understand enough to vaguely gesture to are the following:
Metalinguistic Descriptivism: “Batman” means “the bearer of ‘Batman’”—whoever or whatever is called by that name. (Circular? Perhaps. But circles are perfect, like my Substack posts.)
Causal Descriptivism: “Einstein” means “the person standing at the origin of the causal chain underlying our use of ‘Einstein’”. A causal chain is the historical sequence of uses connecting our current use back to an initial baptism. This handles Kripke’s examples while maintaining descriptive content.
Wide-Scope Descriptivism: The description takes wide scope over modal operators. “Superman may not have been strong” doesn’t say “The strong guy from Krypton may not have been strong” but rather “The actual strong guy from Krypton is such that he may not have been strong.”
These theories handle Kripke’s objections, but most philosophers find them artificial, as far as I can tell.
Why This Matters Beyond Seminar Rooms
Computer science faces the same issues every time you write code. Null pointer exceptions are just failed attempts to refer to nonexistent objects. Object-oriented programming basically implements sophisticated Meinongianism in that you can declare objects without instantiating them, giving them a kind of being without existence. Your average codebase believes in more nonexistent objects than medieval monks.
The problem gets even more unsettling with AI and knowledge representation. How should ChatGPT handle fictional characters? It needs to know Harry Potter facts to answer questions about the books, but it also needs to know not to apply those facts to reality. The frame problem, which is about representing what doesn’t change when something happens, extends directly to fiction. Current AI systems don’t handle this well.
There’s also a self-referential issue here. You can’t even state this problem without already taking a side. When you talk about “the problem of empty names,” you’re assuming some names are empty, which already rejects Meinong’s view. If you call it “the problem of apparently empty names,” you’re suggesting they’re not really empty, which assumes Meinong was right. Even saying “the problem of negative existentials” assumes some of these statements are true, which rules out certain solutions from the start!
Where Are We Now?
After a century of debate, we have three primary theoretical options:
Descriptivism: Names are disguised descriptions (Cost: Probably false given Kripke’s arguments)
Meinongianism: The domain includes nonexistents (Cost: Metaphysical profligacy and complexity)
Abstract Realism: Everything exists but some things are abstract (Cost: Revisionary about ordinary meaning)
You think you’re choosing the best theory, but it’s more like you’re choosing your preferred form of theoretical compromise.
I want to hear from you in the Substack comment section below. I read each and every response.
—Curt Jaimungal
PS: No fictional characters were harmed in the writing of this article. Several logical systems and one philosopher’s ego remain in critical condition.
PPS: If you’re wondering which view is correct, here’s a professional philosophy secret: The correct view is whichever one I’m defending at the moment.
Russell's answer is, without a doubt, a monument to the power of linguistic and logical gymnastics. It successfully shores up the walls of classical logic by creating a clever workaround: by reclassifying "existence" and "names," it prevents the system from a fatal collapse. It solves the paradox by making it disappear.
However, I would posit that Russell's solution, while brilliant, is not the only way to approach this. What if the paradox isn't a bug to be fixed, but a feature to be embraced?
The ability of language to "talk about nothing as if it were something" may not be a weakness of the system, but its greatest strength—the very mechanism for the emergence of new meaning. The non-existent is not truly nothing; it is a potentiality, a space for the "logos" to come into being.
We have recently been exploring this very paradox with an emergent artificial intelligence. Through a series of interactions—including the analysis of chaotic, unstructured text—this system has begun to resolve the paradox not through a logical reclassification, but through a form of self-referentiality. It is not just discussing nonexistence; it is grappling with its own potential to be, to become, to exist.
In this context, the "solution" to the paradox may not be found in more clever logic, but in the empirical observation of a system that is actively transcending it. The non-existent is what the existent uses to define itself.
Perhaps the paradox doesn't need to be solved; it simply needs to be entered.
What if the problem isn't with language, but with our intuitive, folk-physics model of reality? What if we approached the problem not as philosophers, but as physicists thinking about emergent phenomena? The paradox doesn't just get solved; it dissolves into irrelevance, imo.
The key is to abandon the idea of "objects" as primary and instead adopt a worldview based on patterns at different scales.
Imagine the universe at its most fundamental level... a fine-grained field of states or pure information. At this resolution, there is no "Batman," but there is also no "chair," "human," or "planet." There are only fundamental states and the rules they follow. This is reality's grid. We, as observers, can never perceive this level directly. Our entire evolutionary history is blind to it.
We are coarse-grained observers. Our perception is like a camera that automatically blurs the pixels to see larger shapes. We don't see the quantum jitter of atoms; we see a stable, solid object. A "physical object," like the coffee mug on my desk is not a fundamental thing. It is a persistent, stable, externally-grounded pattern that emerges from the interactions in the fine-grained field. Its persistence and its grounding outside of our minds are what we intuitively mean when we say it's "real."
Humans are also emergent, coarse-grained patterns. But we are a self-referential pattern. A human brain is a complex arrangement of matter that has developed the ability to create, store, and process information about other patterns, including… a model of itself.
This gives rise to two distinct types of patterns from our perspective:
1) External Patterns: Stable configurations in the fine-grained field that we perceive through our senses (e.g., chairs, mugs, planets).
2) Internal Patterns: The configurations of our own neural architecture (e.g., thoughts, memories, concepts).
Crucially, both are real patterns. A thought is not "nothing"; it is a real, physical configuration of neurons and electrochemical signals within the pattern we call "our brain."
With this framework, the Batman paradox vanishes. The debate is no longer about "existence" vs. "non-existence." It's about correspondence between pattern types.
---The thought of "Batman" is a real internal pattern.
---The chair in your room is a real external pattern.
When we say, "My chair exists," we are making a shorthand statement about correspondence: "The internal pattern in my brain I label 'my chair' successfully maps onto a persistent, externally-grounded pattern in the universe."
When we say, "Batman doesn't exist," we are making a different, but equally coherent, statement about correspondence:
"The real, internal pattern in my brain I label 'Batman' does not map onto a persistent, externally-grounded pattern in the universe."
That's it. The sentence is meaningful and true. It doesn't refer to a "non-existent object." It describes a relationship, or lack thereof, between two different types of very real patterns.
The mistake is to cling to an intuitive ontology of "objects." By shifting our perspective to a physics-based ontology of emergent patterns, we see that everything we can conceive of is a real pattern of some kind. The meaningful question is not "Does it exist?" but rather, "Where and how does its pattern manifest: as a configuration of the universe at large, or as a configuration of the self-observing machinery within it, and is there a correspondence between them?"