abstract duality: opposite category,
concrete duality: dual object, dualizable object, fully dualizable object, dualizing object
between higher geometry/higher algebra
Langlands duality, geometric Langlands duality, quantum geometric Langlands duality
(0,1)-category theory: logic, order theory
proset, partially ordered set (directed set, total order, linear order)
distributive lattice, completely distributive lattice, canonical extension
natural deduction metalanguage, practical foundations
type theory (dependent, intensional, observational type theory, homotopy type theory)
computational trinitarianism = propositions as types +programs as proofs +relation type theory/category theory
In logic, de Morgan duality is a duality between intuitionistic logic and dual-intuitionistic paraconsistent logic. In classical logic and linear logic, it is a self-duality mediated by negation. Although it goes back to Aristotle (at least), its discovery is generally attributed to Augustus de Morgan.
More explicitly, de Morgan duality is the duality between logical operators as shown in the table below:
Intuitionistic operator | Dual-intuitionistic operator | |
---|---|---|
$\top$ (truth) | $\bot$ (falsehood) | |
$\wedge$ (conjunction) | $\vee$ (disjunction) | |
$\Rightarrow$ (conditional) | $\setminus$ (without?) | |
$\Leftrightarrow$ (biconditional) | $+$ (exclusive disjunction) | |
$\neg$ ($p \Rightarrow \bot$) | $-$ ($\top \setminus p$) | |
$\forall$ (universal quantification) | $\exists$ (existential quantification) | |
$\Box$ (necessity) | $\lozenge$ (possibility) |
The first two operators in each column exist in both intuitionistic and dual-intuitionistic propositional logic and the last two in each column exist in both forms of predicate logic and modal logic (respectively), but they are still dual as shown. All of these exist in classical logic (although some of the paraconsistent operators are not widely used), and the two forms of negation ($\neg$ and $-$) are the same there.
In linear logic, this extends to a duality between conjunctive and disjunctive operators:
Conjunctive operator | Disjunctive operator | |
---|---|---|
$\top$ | $0$ | |
$1$ | $\bot$ | |
$\&$ | $\oplus$ | |
$\otimes$ | $\parr$ | |
$^\bot$ | $^\bot$ | |
$\bigwedge$ | $\bigvee$ | |
$!$ | $?$ |
As with classical negation, linear negation is self-dual. (For the categorical semantics of this see at dualizing object and at Wirthmüller context – Comparison of push-forwards.)
The first two rows of the intuitionistic/dual-intuitionistic/classical duality generalise to arbitrary lattices, including subobject lattices in coherent categories, and from there to the duality between limits and colimits in category theory:
Limit | Colimit | |
---|---|---|
top | bottom | |
meet | join | |
intersection | union | |
terminal object | initial object | |
⋮ | ⋮ |
So in a way, all duality in category theory is a generalisation of de Morgan duality.
The de Morgan laws are the statements, valid in various forms of logic, that de Morgan duality is mediated by negation. For example, using the second line of the first table, we have
Traditionally, the term is reserved for this line.
In the foundations of constructive mathematics, de Morgan's Law usually means the statement
since every other aspect of the first two lines is already constructively valid, the claim that negation mediates the de Morgan self-duality already has a name (the double negation law, equivalent to the principle of excluded middle), and no other line involves only operators that appear in intuitionstic propositional calculus.