Key Logical Paradoxes

Why This Matters

Paradoxes aren't just clever puzzles. They're stress tests for logical systems. When you encounter a paradox in Formal Logic, you're seeing exactly where rules of inference, truth conditions, and definitions break down. These breakdowns forced logicians to develop more rigorous foundations, from axiomatic set theory to multi-valued logics. Understanding why each paradox creates a contradiction teaches you more about logical structure than memorizing a hundred valid argument forms.

You're being tested on your ability to identify the mechanism behind each contradiction: Is it self-reference? Vagueness? Unrestricted set formation? Don't just memorize the paradoxes. Know what principle each one challenges and how logicians have responded. When an exam question asks you to analyze a novel paradox, you'll need to recognize which family it belongs to and what logical tools address it.

---

Self-Reference Paradoxes

These paradoxes arise when a statement or definition refers back to itself, creating a loop that generates contradiction. The core mechanism is that self-reference allows a statement to assert something about its own truth value, and that assertion collapses under evaluation.

Liar Paradox

The canonical example of semantic self-reference: "This statement is false."

Try assigning it a truth value. If you assume it's true, then what it says holds, so it's false. If you assume it's false, then it's not the case that it's false, so it's true. Every assignment contradicts itself.

• This is sometimes called a dialetheic situation because the statement seems to demand both truth values simultaneously.
• It directly challenges bivalence, the principle that every proposition is either true or false.
• Tarski's response was to separate object language from metalanguage: a sentence in language $L$ can only have its truth predicate defined in a metalanguage $L'$, blocking the self-referential loop.

Epimenides Paradox

The historical ancestor of liar-type reasoning. Epimenides the Cretan declares: "All Cretans are liars."

This is weaker than the Liar because it's only genuinely paradoxical under narrow conditions. If even one other Cretan sometimes tells the truth, Epimenides' statement is simply false and he's just a liar himself. The contradiction only arises if he's the sole Cretan, or if every other Cretan always lies. It demonstrates contingent self-reference, where context determines whether a contradiction actually follows.

Curry's Paradox

Consider the sentence: "If this statement is true, then $P$", where $P$ is any arbitrary proposition (say, "the moon is made of cheese").

Here's why it's devastating:

1. Call the sentence $C$. So $C$ says: "If $C$ is true, then $P$."
2. Suppose $C$ is true. Then the conditional "If $C$ is true, then $P$" holds.
3. Since we're assuming $C$ is true, and the conditional holds, by modus ponens we get $P$.
4. So we've shown: if $C$ is true, then $P$. But that's exactly what $C$ says.
5. Therefore $C$ is true. And by step 3, $P$ follows.

This proves any arbitrary proposition $P$. Unlike the Liar, which creates a local contradiction, Curry's Paradox trivializes the entire system by making everything provable. It exploits the interaction between self-reference and the logical properties of the material conditional.

---

Set-Theoretic Paradoxes

These paradoxes expose problems with unrestricted set formation, the naive assumption that any definable collection forms a legitimate set. The mechanism is that certain definitions generate sets whose membership conditions are self-contradictory.

Russell's Paradox

Define the set $R = \{x \mid x \notin x\}$, the set of all sets that don't contain themselves. Now ask: does $R \in R$?

• If $R \in R$, then by definition $R$ doesn't contain itself, so $R \notin R$.
• If $R \notin R$, then $R$ meets its own membership condition, so $R \in R$.

Either way, contradiction. This destroyed naive set theory, which assumed any well-defined property determines a set (the unrestricted comprehension axiom). The response was axiomatic set theories like ZFC, which restrict comprehension so you can only form subsets of already-existing sets, blocking $R$ from being constructed in the first place.

This is the most historically significant paradox in the foundations of mathematics. Know it thoroughly.

Barber Paradox

An informal illustration of the same logical structure: a barber in a town shaves exactly those residents who don't shave themselves. Who shaves the barber?

The resolution here is simpler than Russell's. You can just deny that such a barber exists. No logical system collapses; the scenario is simply an impossible description, like "a square circle." The key difference is that in naive set theory, $R$ must exist because the comprehension axiom guarantees it. The barber has no such guarantee.

---

Semantic and Definability Paradoxes

These paradoxes arise from self-referential definitions involving language, naming, or description. The mechanism involves definitions that reference the very system of definition itself.

Berry Paradox

Consider: "The smallest positive integer not definable in fewer than eleven words."

That phrase itself is a definition, and it contains only ten words. So it defines the number in fewer than eleven words, contradicting the claim that the number isn't definable in fewer than eleven words.

• The paradox targets definability rather than truth. It shows that "definable in language $L$" can't be consistently applied within $L$ itself.
• This connects directly to Gödel's incompleteness results and Tarski's undefinability theorem, both of which establish limits on what formal systems can express about themselves.

Grelling–Nelson Paradox

Some adjectives describe themselves: "short" is a short word, "English" is an English word. Call these autological. Others don't: "long" isn't a long word, "German" isn't a German word. Call these heterological.

Now ask: is "heterological" heterological?

• If yes, it describes itself, which means it's autological, not heterological. Contradiction.
• If no, it doesn't describe itself, which means it is heterological. Contradiction.

This has a parallel structure to Russell's Paradox but operates entirely in natural language rather than set theory. Where Russell asks about set membership, Grelling–Nelson asks about predicate self-application.

---

Vagueness Paradoxes

These paradoxes exploit the absence of sharp boundaries in ordinary predicates. The mechanism is that tolerance principles ("small differences can't matter") combined with transitivity lead to absurd conclusions.

Sorites Paradox

One grain of sand isn't a heap. And adding a single grain to something that isn't a heap can't make it a heap. But apply that reasoning 10,000 times and you're forced to conclude that 10,000 grains isn't a heap either.

The tolerance premise ("one grain can't make the difference") seems obviously true for any individual step, yet iterated application produces an obviously false conclusion. This challenges the law of excluded middle for vague predicates, since there's no determinate grain where "not a heap" becomes "a heap."

Responses include:

• Fuzzy logic, which assigns truth values on a continuous scale (e.g., 0.0 to 1.0) rather than just true/false
• Supervaluationism, which says borderline cases are neither true nor false but preserves classical tautologies
• Epistemicism, which insists there is a sharp boundary but we can't know where it is

---

Epistemic and Modal Paradoxes

These paradoxes involve knowledge, expectation, and reasoning about what agents can know or predict. The mechanism typically involves self-defeating predictions or knowledge that undermines itself.

Unexpected Hanging Paradox

A prisoner is told he'll be hanged on one day next week, and that he won't be able to predict which day it is the morning before. He reasons by backward induction: it can't be Friday (he'd know Thursday night), so it can't be Thursday (Friday's eliminated, so he'd know Wednesday night), and so on through every day. He concludes the hanging is impossible.

Then he's hanged on Wednesday and is genuinely surprised. His "proof" was flawed.

The paradox challenges assumptions about the interaction of knowledge and surprise. The prisoner's reasoning seems valid at each step, yet the conclusion is clearly wrong. Pinpointing exactly where it fails requires careful epistemic logic, and there's still no consensus on the best resolution.

Zeno's Paradoxes

Achilles and the Tortoise: Achilles gives the tortoise a head start. To catch up, he must first reach where the tortoise was, but by then the tortoise has moved further. He must reach that point, and so on. Infinitely many tasks seem required.

The Dichotomy: To travel any distance, you must first travel half of it, then half of the remainder, and so on ad infinitum. You seemingly can never finish.

The mathematical resolution uses convergent infinite series. For instance, $\frac{1}{2} + \frac{1}{4} + \frac{1}{8} + \cdots = 1$. An infinite number of steps can sum to a finite quantity, so the task is completable. Still, philosophical debate continues about whether the mathematical solution fully addresses the metaphysical question of how infinitely many sub-tasks can be completed in finite time.

---

Quick Reference Table

---

Self-Check Questions

1. Russell's Paradox and the Barber Paradox share the same logical structure but differ in whether they threaten formal systems. What makes one "unavoidable" under naive set theory and the other "merely impossible"?

2. If asked to identify a paradox that challenges bivalence through vagueness rather than self-reference, which paradox would you choose, and what makes it structurally different from the Liar?

3. Both Curry's Paradox and the Liar Paradox involve self-reference. Why is Curry's considered more dangerous to logical systems? What's the difference between local contradiction and global triviality?

4. Restricting set comprehension axioms solves Russell's Paradox. Explain why the same strategy wouldn't help with the Sorites Paradox. What kind of solution does Sorites require instead?

5. Which paradoxes would you cite as evidence that natural language contains features that resist formalization? What specific mechanism does each exploit?