Paraconsistent Logics

A database contains: 'Customer A has paid' and 'Customer A has not paid' (data entry error). In classical logic, from this follows... anything, including 'The moon is made of cheese'. The system becomes useless. Paraconsistent logics allow isolating the contradiction and continuing to work.

• **Knowledge bases**: merging data from different sources often creates contradictions. The system must continue answering queries
• **Legal systems**: laws can contradict each other. A court must reach a verdict despite the conflict
• **AI assistants**: training data contains contradictory information. The model must give meaningful responses

Paraconsistent Logic

**Paraconsistent logic** - a family of logics in which the presence of a contradiction **does not destroy the entire system**. In classical logic, anything follows from a contradiction (ex contradictione quodlibet). Paraconsistent logics block this inference.

The name: para- (beside, near) + consistency. A logic 'near consistency' - tolerates contradictions but doesn't collapse.

Why do we need such a logic? In reality, we constantly work with contradictory information:

Contradiction Tolerance

**Contradiction tolerance** - the key property of paraconsistent logics. A system can contain both A and ¬A simultaneously, but this does not obligate accepting any arbitrary B.

Comparing classical and paraconsistent logics:

**Localization principle**: a contradiction in one part of a knowledge base should not infect the whole system. Paraconsistent logics ensure this isolation.

Which inference rule is modified? The main target is **disjunctive syllogism**: A ∨ B, ¬A ⊢ B. In paraconsistent logics, it is restricted or disabled.

The Explosion Problem

**Explosion** - the classical logic principle that anything follows from a contradiction. The Latin name is **ex contradictione quodlibet** (ECQ) or **ex falso quodlibet**.

Why is explosion a problem? In real reasoning, contradictions arise from errors, incomplete information, and conflicting sources. If every contradiction makes a system useless, we cannot work with reality.

**Belnap's four-valued logic** is popular in CS: a value can be 'known true', 'known false', 'contradictory' (both), or 'unknown' (neither). This is natural for knowledge bases.

Dialethism

**Dialethism** - the philosophical position that some contradictions are **true**: there exist propositions that are simultaneously true and false (true contradictions, dialetheia).

The main advocate of dialethism is Graham Priest. His Logic of Paradox (LP) is a formal system in which dialetheia are logically permissible.

Dialethism is a radical position. Most logicians reject true contradictions but accept paraconsistency as a tool for working with imperfect information.

Applications: automated reasoning with contradictory knowledge bases, revision theory of truth, formalization of paradoxes, legal reasoning under conflicting norms.

Key Ideas

• **Paraconsistency** - the ability of a system to tolerate contradictions without collapsing
• **Explosion (ECQ)**: A ∧ ¬A ⊢ B - anything follows from a contradiction. Paraconsistent logics block this
• **Dialethism** - the radical thesis that some contradictions are true (true contradictions)
• Practice: LP, Belnap's four-valued logic, relevance logic - tools for working with contradictory information

Related Topics

Paraconsistency is connected to other non-classical logics:

• Liar Paradox — The main motivating example for dialethism
• Modal Logic — Paraconsistent modal logics explore 'possible contradictions'

Вопросы для размышления

• If a database contains a contradiction, what is better: refuse to answer, give both answers, or choose one? Why?
• Can you think of a situation where accepting a true contradiction would be more useful than eliminating it?
• How does 'contradiction in data' differ from 'contradiction in reality'? Does this distinction make sense?

Связанные уроки

• fl-01-intro