nLab Phoa's principle

Context

Computability

Topology

topology (point-set topology, point-free topology)

see also differential topology, algebraic topology, functional analysis and topological homotopy theory

Introduction

Basic concepts

Universal constructions

Extra stuff, structure, properties

Examples

Basic statements

Theorems

Analysis Theorems

topological homotopy theory

(0,1)(0,1)-Category theory

(,1)(\infty,1)-Category theory

Contents

Idea

Phoa’s principle is a principle used to axiomatize Sierpinski space in synthetic topology and synthetic domain theory. Phoa’s principle is also called the Phoa principle in Bauer & Taylor 2009.

Phoa’s principle is named after Wesley Phoa.

Definition

Let LL be a 01-bounded semilattice, that is, a semilattice with an absorbing element. Phoa’s principle states that every endofunction f:LLf:L \to L is monotonic and the precomposition function hhi:L LL 𝟚h \mapsto h \circ i:L^L \to L^\mathbb{2} by the embedding i:𝟚Li:\mathbb{2} \hookrightarrow L of the booleans into LL is an embedding.

If LL is a distributive lattice, then Phoa’s principle is equivalent to the linear interpolation condition that for all endofunctions f:LLf:L \to L and elements xLx \in L:

f(x)=f()(xf())f(x) = f(\top) \wedge (x \vee f(\bot))

Properties

There are many types of distributive lattices for which non-trivial such distributive lattices cannot satisfy Phoa’s principle.

Theorem

Suppose LL is a distributive lattice LL with an endpoint-swapping endofunction ¬:LL\neg:L \to L, where ¬()=\neg(\top) = \bot and ¬()=\neg(\bot) = \top. Then the only such distributive lattice LL for which Phoa’s principle holds is the trivial Heyting algebra.

Proof

Applying the function x¬xx \mapsto \neg x to the linear interpolation condition for the distributive lattice yields

¬x=¬()(x¬)=(x)=\neg x = \neg (\top) \wedge (x \vee \neg \bot) = \bot \wedge (x \vee \top) = \bot

Substituting \bot into the equation yields

¬==\neg \bot = \top = \bot

which is precisely the condition that LL be a trivial such distributive lattice.

Lemma

Suppose LL is a Heyting algebra for which Phoa’s principle holds. Then LL is the trivial Heyting algebra.

Proof

LL being a Heyting algebra implies that LL has a Heyting implication x,yxyx, y \mapsto x \to y and thus a Heyting negation function x¬xx \mapsto \neg x defined by ¬xx\neg x \coloneqq x \to \bot such that ¬()=\neg(\top) = \bot and ¬()=\neg(\bot) = \top. Hence, by the first theorem, LL is the trivial Heyting algebra if Phoa’s principle holds.

Theorem

Suppose LL is a De Morgan algebra for which Phoa’s principle holds. Then LL is the trivial De Morgan algebra.

Proof

LL being a De Morgan algebra implies that LL has an involution x¬xx \mapsto \neg x such that ¬()=\neg(\top) = \bot and ¬()=\neg(\bot) = \top. Hence, by the first theorem, LL is the trivial De Morgan algebra if Phoa’s principle holds.

Lemma

Suppose LL is a finite distributive lattice for which Phoa’s principle holds. Then LL is the trivial distributive lattice.

Proof

If LL is a finite distributive lattice, then there are many such functions f:LLf:L \to L such that f()=f(\top) = \bot and f()=f(\bot) = \top definable using the universal property of the finite set with nn elements. Hence, by the first theorem, LL is the trivial distributive lattice if Phoa’s principle holds.

Lemma

There are no distributive lattices LL, whose carrier set is the set of natural numbers, for which Phoa’s principle holds.

Proof

Suppose that LL is a distributive lattice whose carrier set is the set of natural numbers. Then there are many such functions f:LLf:L \to L such that f()=f(\top) = \bot and f()=f(\bot) = \top definable using the universal property of the natural numbers. Hence, by the first theorem, LL is the trivial distributive lattice if Phoa’s principle holds, which contradicts that the carrier set of LL is the set of natural numbers.

Examples

References

Last revised on August 26, 2026 at 14:23:29. See the history of this page for a list of all contributions to it.