What is Existence, Exactly? Summary — Theories of Everything

This episode explores the philosophical problem of existence through the lens of fictional characters like Batman, tracing how major thinkers have tried to resolve the paradox that we can meaningfully talk about things that don’t exist.

Saying “Batman doesn’t exist” creates a logical puzzle: the name “Batman” appears to refer to something, yet the sentence denies there is anything to refer to.
This is a negative singular existential — a sentence that denies a particular individual exists — and it seems to require either referring to nothing (like pointing at empty space) or predicating non-existence of a referent that by definition has none.
The puzzle extends to Santa Claus, Pegasus, and other fictional entities: if these sentences are true and meaningful, language can talk about nothing as if it were something.

Bertrand Russell proposed that existence is not a property of individuals (first-order) but a second-order property of properties — specifically, the property of being instantiated.
“Dragons exist” means the property “being a dragon” is instantiated, not that some individual has the property of existence.
Names like “Batman” are not genuine singular terms but disguised definite descriptions (e.g., “the caped crusader of Gotham”).
“Batman doesn’t exist” then translates to: it is not the case that there exists exactly one caped crusader — eliminating singular reference to Batman entirely.
This solution dominated philosophy for decades until Saul Kripke’s Naming and Necessity in the 1970s.

Kripke showed names and descriptions behave differently in modal contexts (involving possibility and necessity).
“Peter Parker may not have been bitten by a radioactive spider” is true; “the guy bitten by a radioactive spider may not have been bitten by a radioactive spider” is false — so the name cannot be synonymous with the description.
Names are rigid designators: they pick out the same individual in every possible world where that individual exists; descriptions are non-rigid and can pick out different individuals across worlds.
One can refer to Feynman while believing false descriptions about him (e.g., that he invented the hula hoop) — the name points directly to the bearer, not via associated descriptions.
This revived the original problem: if “Batman” is a rigid designator, what does it designate when there is no Batman?

Alexius Meinong proposed that the domain of quantification includes objects that don’t exist — existence is a genuine first-order property that only some objects have.
Comprehension principle: for any set of properties, there is an object having exactly those properties; some such objects exist (a USB-C cord), some don’t (Batman).
“Batman doesn’t exist” is then straightforward: the object designated by “Batman” lacks the property of existence.
Russell objected that this leads to contradictions: a round square (violating non-contradiction), a tall man with no facial hair status (violating excluded middle), and an existent golden mountain (thinking something into existence).

Modern Meinongians distinguish nuclear properties (ordinary characterizing properties like being red, tall, a detective) from extra-nuclear properties (higher-order metaphysical/logical properties like existing, being possible, being impossible).
Comprehension applies only to nuclear properties, blocking the existent golden mountain (existence is extra-nuclear) and the round square (impossibility is extra-nuclear).
Edward Zalta’s dual copula strategy: objects can encode properties (have them as part of their nature, like fictional characters) or exemplify them (instantiate them in reality).
Sherlock Holmes encodes being a detective but exemplifies being fictional; the round square encodes roundness and squareness without exemplifying either — avoiding contradiction.
These solutions work but require complex bookkeeping of property types and predication modes.

A recent approach: all genuine singular terms refer, everything exists, but apparently empty names like “Batman” refer to abstract artifacts created by humans (e.g., by DC Comics).
Batman exists as an abstract object (like numbers or stories); “Batman doesn’t exist” expresses the true proposition that Batman isn’t concrete.
This preserves direct reference (Kripke) and classical logic while avoiding Meinongian non-existent objects.
Cost: it’s revisionary about ordinary usage — when people say “Santa doesn’t exist,” they typically mean he doesn’t exist in any manner, not just that he’s non-concrete.

Some philosophers defend descriptivism with modifications:
- Metalinguistic descriptivism: “Batman” means the bearer of the name “Batman” (potentially circular).
- Causal descriptivism: “Einstein” means the person at the origin of the causal chain of uses linking back to an initial baptism.
- Wide-scope descriptivism: descriptions take wide scope over modal operators — “Superman may not have been strong” means the actual strong guy from Krypton is such that he may not have been strong.
Most philosophers find these constructions artificial.

Null pointer exceptions are failed attempts to refer to non-existent objects — the same problem in code.
Object-oriented programming implements a kind of Meinongianism: classes declare objects without instantiating them (being without existence).
AI systems like ChatGPT must represent fictional knowledge (Harry Potter facts) without applying it to reality — an instance of the frame problem extended to fiction.
Current AI systems handle this poorly, though they are evolving.

You cannot state the problem without presupposing a solution:
- “Problem of empty names” assumes some names are empty (rejects Meinong).
- “Problem of apparently empty names” suggests they aren’t really empty (assumes Meinong).
- “Problem of negative existentials” assumes some such statements are true (rules out certain solutions).

Descriptivism (names as disguised descriptions): cost is Kripke’s arguments seem to falsify it; requires complex counter-moves.
Meinongianism (domain includes non-existent objects): cost is metaphysical profligacy/fecundity and technical complexity (nuclear/extra-nuclear, encoding/exemplifying).
Abstract realism (everything exists, some things abstract): cost is revisionary — doesn’t match what ordinary speakers mean by “existence.”
Choosing a theory is less about finding the correct one than selecting a preferred form of theoretical compromise.

The discussion so far treats existence linguistically (how it functions in sentences), but there may be more to existence than linguistic relations.
Upcoming topics: Heidegger on what existence is; Leibniz on “why something rather than nothing”; van Inwagen on whether existence is a single thing; presentism (only the present exists) vs. block universe views on temporal ontology.