Alphabeta Math
Session-authored (Fable 5 assisted)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced — the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted — a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated — a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

10 results · all verified · 0 also independently AI-judged
Every result on this page is machine-checked by a proof checker and read in full by a delegated reviewing agent on the owner's instruction; the judge is an additional, independent cross-model AI review of the proofs. The 10 not AI-judged were verified by owner audit (typically over a confirmed judge false positive), not failures.

Relations, Functions, and Quotients: Examples and Counterexamples

1 · Prerequisites

2 · Summary

3 · Logical flowchart

4 · Definitions, theorems and proofs

None yet.

5 · Examples, counterexamples and false statements

ExampleConstruction: AI-adaptedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

A four-pair relation on {,{},{,{}}}\{\varnothing,\{\varnothing\},\{\varnothing,\{\varnothing\}\}\} with its domain, range, field, inverse, restriction, image and self-composite computed

Example

Write u:=u := \varnothing, v:={}v := \{\varnothing\} and w:={,{}}w := \{\varnothing,\{\varnothing\}\}; these are pairwise distinct, since uu has no element, vv has exactly the element uu, and ww has vv as an element while vv does not. Put A:={u,v}{w}A := \{u,v\} \cup \{w\} and

R:={(u,v),(v,w)}{(w,u),(u,u)}.R := \{(u,v),(v,w)\} \cup \{(w,u),(u,u)\}.

Then RR is a relation on AA with

domR=ranR=fldR=A,R1={(v,u),(w,v)}{(u,w),(u,u)},\operatorname{dom} R = \operatorname{ran} R = \operatorname{fld} R = A, \qquad R^{-1} = \{(v,u),(w,v)\} \cup \{(u,w),(u,u)\},

R{u}={(u,v),(u,u)},R[{u}{v}]=A,RR={(u,w),(u,v)}{(u,u),(v,u)}{(w,v),(w,u)}.R \restriction \{u\} = \{(u,v),(u,u)\}, \qquad R[\{u\} \cup \{v\}] = A, \qquad R \circ R = \{(u,w),(u,v)\} \cup \{(u,u),(v,u)\} \cup \{(w,v),(w,u)\}.

Facts & Assumptions

Given: u:=u := \varnothing, v:={}v := \{\varnothing\}, w:={,{}}w := \{\varnothing,\{\varnothing\}\}, A:={u,v}{w}A := \{u,v\} \cup \{w\} and RR as displayed.

[L2]

domR:={a:b (a,b)R},ranR:={b:a (a,b)R}\operatorname{dom} R := \{\, a : \exists b\ (a,b) \in R \,\}, \qquad \operatorname{ran} R := \{\, b : \exists a\ (a,b) \in R \,\} (Relation, domR\operatorname{dom} R, ranR\operatorname{ran} R, fldR\operatorname{fld} R, and the specialisations "relation from AA to BB" and "relation on AA").

[L3]

(b,a)R1(b,a) \in R^{-1} holds if and only if (a,b)R(a,b) \in R; (a,c)SR(a,c) \in S \circ R holds if and only if (a,b)R(a,b) \in R and (b,c)S(b,c) \in S for some bb; and (a,b)RA(a,b) \in R \restriction A holds if and only if (a,b)R(a,b) \in R and aAa \in A (The inverse relation R1R^{-1}, the composite SRS \circ R, and the restriction RAR \restriction A).

[L4]

bR[A]b \in R[A] holds if and only if (a,b)R(a,b) \in R for some aAa \in A (The image R[A]R[A] and the preimage R1[B]R^{-1}[B] of a set under a relation).

[L5]

(a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d ((a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d).

[L6]

(a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\} (The Kuratowski ordered pair (a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\}).

[L7]

{x,y}\{x,y\} is the set whose elements are exactly xx and yy, and {x}:={x,x}\{x\} := \{x,x\} (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

[L8]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

Verification

technique · direct
1.1

uu, vv and ww are pairwise distinct and AA has exactly them as elements; every element of RR is an ordered pair with both coordinates in AA, so RR is a relation on AA, that is RA×AR \subseteq A \times A.

L1L6L7L8L9L10L12
2.1

Reading off first coordinates gives uu, vv, ww, so domR=A\operatorname{dom} R = A; reading off second coordinates gives vv, ww, uu, so ranR=A\operatorname{ran} R = A; and the field, being the union of the two, is AA as well.

L2L5L9L11step 1.1
2.2

Reversing each pair gives R1={(v,u),(w,v)}{(u,w),(u,u)}R^{-1} = \{(v,u),(w,v)\} \cup \{(u,w),(u,u)\}; keeping the pairs whose first coordinate is uu gives R{u}={(u,v),(u,u)}R \restriction \{u\} = \{(u,v),(u,u)\}; and the elements related to uu or to vv are vv, uu and ww, so R[{u}{v}]=AR[\{u\} \cup \{v\}] = A.

L3L4L5L7L9L11step 1.1
2.3

For the self-composite, each pair (a,b)(a,b) of RR is matched with every pair of RR whose first coordinate is bb: from (u,v)(u,v) and (v,w)(v,w) comes (u,w)(u,w); from (u,u)(u,u) and the two pairs starting at uu come (u,v)(u,v) and (u,u)(u,u); from (v,w)(v,w) and (w,u)(w,u) comes (v,u)(v,u); and from (w,u)(w,u) and the two pairs starting at uu come (w,v)(w,v) and (w,u)(w,u). These six pairs are exactly the displayed RRR \circ R.

L3L5L9L11step 1.1
3.1

The domain, range, field, inverse, restriction, image and self-composite are as displayed.

step 2.1step 2.2step 2.3
ExampleConstruction: Literature-sourcedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

\varnothing is a relation on every set, is the unique equivalence relation on \varnothing, is a function B\varnothing \to B for every BB, is a bijection \varnothing \to \varnothing, and is not a surjection {}\varnothing \to \{\varnothing\}

Example

The empty set does the work of five different objects at once.

  • \varnothing is a relation, and a relation on AA for every set AA; its domain, range and field are all \varnothing.
  • \varnothing is the only relation on \varnothing, and it is an equivalence relation on \varnothing; so \varnothing carries exactly one equivalence relation.
  • \varnothing is a function B\varnothing \to B for every set BB, and it is the only one.
  • \varnothing is a bijection \varnothing \to \varnothing.
  • \varnothing is not a surjection {}\varnothing \to \{\varnothing\}, even though it is an injective function {}\varnothing \to \{\varnothing\}.

The last two together are the reason a codomain belongs to the declaration f:ABf : A \to B rather than to the set ff: one and the same set is a bijection under one declaration and a non-surjection under another. The empty function is also the unique element of the empty product.

Facts & Assumptions

Given: the set \varnothing and arbitrary sets AA and BB.

[L2]

domR:={a:b (a,b)R},ranR:={b:a (a,b)R}\operatorname{dom} R := \{\, a : \exists b\ (a,b) \in R \,\}, \qquad \operatorname{ran} R := \{\, b : \exists a\ (a,b) \in R \,\} (Relation, domR\operatorname{dom} R, ranR\operatorname{ran} R, fldR\operatorname{fld} R, and the specialisations "relation from AA to BB" and "relation on AA").

[L3]
[L4]

We write f:ABf : A \to B, and say ff is a function from AA to BB, when ff is a function with domf=A\operatorname{dom} f = A and ranfB\operatorname{ran} f \subseteq B (A function is a relation ff with (a,b)f(a,b) \in f and (a,c)f(a,c) \in f implying b=cb = c; f:ABf : A \to B, the value f(a)f(a), domain and codomain).

[L5]

ff is injective (one-to-one) if f(x)=f(y)f(x) = f(y) implies x=yx = y, for all x,yAx, y \in A (Injection, surjection, bijection).

[L6]

ff is surjective (onto) if for every bBb \in B there is some xAx \in A with f(x)=bf(x) = b (Injection, surjection, bijection).

[L7]

A binary relation \sim on AA is an equivalence relation when it is reflexive, symmetric and transitive (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L8]

RR is reflexive on AA when (a,a)R(a,a) \in R for every aAa \in A (Reflexive, irreflexive, symmetric, asymmetric, antisymmetric, transitive, and connex relations on a set).

[L9]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

[L10]

{x}:={x,x}\{x\} := \{x,x\}, the singleton of xx, is the set whose only element is xx (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

Verification

technique · direct
1.1

\varnothing has no elements, so "every element is an ordered pair" holds vacuously and \varnothing is a relation; A×A\varnothing \subseteq A \times A for every AA, so it is a relation on every set; and no set satisfies the defining conditions for its domain or its range, so both are \varnothing, hence so is its field.

L1L2L9L12
1.2

A relation on \varnothing is a subset of ×\varnothing \times \varnothing, which is \varnothing, so \varnothing is the only one. It is reflexive on \varnothing, symmetric and transitive, since each condition quantifies over elements of \varnothing; hence it is the unique equivalence relation on \varnothing.

L7L8L9L11L12L13
2.1

\varnothing is single valued vacuously, has domain \varnothing and range B\varnothing \subseteq B, so :B\varnothing : \varnothing \to B for every BB; and any function with domain \varnothing has no elements, so it is \varnothing.

L3L4L9L12step 1.1
3.1

As a function \varnothing \to \varnothing it is injective, since the injectivity condition quantifies over elements of the domain, and surjective, since the surjectivity condition quantifies over elements of the codomain and \varnothing has none; so it is a bijection.

L5L6L9step 2.1
3.2

As a function {}\varnothing \to \{\varnothing\} it is still injective, for the same reason, but not surjective: \varnothing is an element of {}\{\varnothing\} and no element of the domain is sent to it.

L5L6L9L10step 2.1
3.3

The empty function is the unique element of the empty product: iAi={}\prod_{i \in \varnothing} A_i = \{\varnothing\}, and its one element is a function with domain \varnothing.

L14L15step 2.1
4.1

All five descriptions hold of the single set \varnothing, and the last two differ only in the declared codomain.

step 1.1step 1.2step 2.1step 3.1step 3.2step 3.3
CounterexampleConstruction: AI-adaptedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

Relations RR and SS on a two-element set with SRRSS \circ R \neq R \circ S

Statement refuted

Refuted claim: SR=RSS \circ R = R \circ S for all relations RR and SS. With u:=u := \varnothing, v:={}v := \{\varnothing\}, A:={u,v}A := \{u,v\} and

R:={(u,v)},S:={(v,u)},R := \{(u,v)\}, \qquad S := \{(v,u)\},

the two composites are SR={(u,u)}S \circ R = \{(u,u)\} and RS={(v,v)}R \circ S = \{(v,v)\}, which are different. Composition of relations is associative (T(SR)=(TS)RT \circ (S \circ R) = (T \circ S) \circ R, (SR)1=R1S1(S \circ R)^{-1} = R^{-1} \circ S^{-1}, (R1)1=R(R^{-1})^{-1} = R, dom(R1)=ranR\operatorname{dom}(R^{-1}) = \operatorname{ran} R, and ΔBR=R=RΔA\Delta_B \circ R = R = R \circ \Delta_A for a relation RR from AA to BB) but not commutative.

Facts & Assumptions

Given: u:=u := \varnothing, v:={}v := \{\varnothing\}, A:={u,v}A := \{u,v\}, R:={(u,v)}R := \{(u,v)\} and S:={(v,u)}S := \{(v,u)\}.

[L1]

(a,c)SR(a,c) \in S \circ R holds if and only if (a,b)R(a,b) \in R and (b,c)S(b,c) \in S for some bb (The inverse relation R1R^{-1}, the composite SRS \circ R, and the restriction RAR \restriction A).

[L3]

(a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d ((a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d).

[L4]

(a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\} (The Kuratowski ordered pair (a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\}).

[L5]

{x,y}\{x,y\} is the set whose elements are exactly xx and yy, and {x}:={x,x}\{x\} := \{x,x\} (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

[L6]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

Counterexample

technique · direct
1.1

uvu \neq v, because vv has an element and uu has none; and RR and SS are relations on AA, each with one pair, both of whose coordinates lie in AA.

L2L5L6L9
2.1

SRS \circ R: the only pair of RR is (u,v)(u,v), and the only pair of SS whose first coordinate is vv is (v,u)(v,u), so the only element of SRS \circ R is (u,u)(u,u).

L1L3L5L7step 1.1
2.2

RSR \circ S: the only pair of SS is (v,u)(v,u), and the only pair of RR whose first coordinate is uu is (u,v)(u,v), so the only element of RSR \circ S is (v,v)(v,v).

L1L3L5L7step 1.1
3.1

(u,u)(u,u) and (v,v)(v,v) are different, since equality of ordered pairs would force u=vu = v; so the two composites have different elements and SRRSS \circ R \neq R \circ S, while the associativity of composition is unaffected.

L3L4L7L8step 1.1step 2.1step 2.2
ExampleConstruction: AI-adaptedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

A two-cell partition of a three-element set, the equivalence relation it induces listed pair by pair, and the quotient set recovered from it

Example

Write u:=u := \varnothing, v:={}v := \{\varnothing\} and w:={,{}}w := \{\varnothing,\{\varnothing\}\}, which are pairwise distinct, and put A:={u,v}{w}A := \{u,v\} \cup \{w\}. The collection {{u,v},{w}}\{\{u,v\},\{w\}\} has nonempty cells, covers AA, and its two cells are disjoint. The equivalence relation it induces on AA, namely "aa and bb lie in the same cell", is

  =  {(u,u),(u,v)}{(v,u),(v,v)}{(w,w)},\sim \;=\; \{(u,u),(u,v)\} \cup \{(v,u),(v,v)\} \cup \{(w,w)\},

with classes [u]=[v]={u,v}[u] = [v] = \{u,v\} and [w]={w}[w] = \{w\}, so that A/={{u,v},{w}}A/{\sim} = \{\{u,v\},\{w\}\} is the original collection again.

Facts & Assumptions

Given: u:=u := \varnothing, v:={}v := \{\varnothing\}, w:={,{}}w := \{\varnothing,\{\varnothing\}\}, A:={u,v}{w}A := \{u,v\} \cup \{w\}, and \sim as displayed.

[L1]

reflexive: aaa \sim a for every aAa \in A (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L2]

symmetric: aba \sim b implies bab \sim a, for all a,bAa, b \in A (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L3]

transitive: aba \sim b and bcb \sim c imply aca \sim c, for all a,b,cAa, b, c \in A (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L4]

[a]  :=  {bA  :  ab}    A[a] \;:=\; \{\, b \in A \;:\; a \sim b \,\} \;\subseteq\; A (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L5]

A/  :=  {[a]  :  aA}A/{\sim} \;:=\; \{\, [a] \;:\; a \in A \,\} (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L7]

{x,y}\{x,y\} is the set whose elements are exactly xx and yy, and {x}:={x,x}\{x\} := \{x,x\} (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

[L8]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

[L11]

(a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d ((a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d).

[L12]

(a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\} (The Kuratowski ordered pair (a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\}).

Verification

technique · direct
1.1

uu, vv and ww are pairwise distinct: uu has no element, vv has exactly uu, and ww has vv as an element while vv does not. So AA has exactly those three elements, and \sim is a relation on AA, each of its five listed pairs having both coordinates in AA.

L7L8L9L10L12L14L15
2.1

\sim is an equivalence relation on AA: it contains (u,u)(u,u), (v,v)(v,v) and (w,w)(w,w), so it is reflexive on AA; its pairs with distinct coordinates are (u,v)(u,v) and (v,u)(v,u), and each of those has its reverse present, so it is symmetric; and every composable pair of its members has its composite present, since chains through uu and vv stay inside {u,v}\{u,v\} and ww relates only to itself.

L1L2L3L11L13step 1.1
3.1

The classes are as stated: [u]={bA:ub}={u,v}[u] = \{b \in A : u \sim b\} = \{u,v\}, [v]={u,v}[v] = \{u,v\} by the pairs (v,u)(v,u) and (v,v)(v,v), and [w]={w}[w] = \{w\}. Hence A/A/{\sim} has exactly the two elements {u,v}\{u,v\} and {w}\{w\}, which is the collection we started from.

L4L5L7L11L13step 1.1step 2.1
4.1

The collection is therefore a partition of AA in the sense of the classes being nonempty, covering AA and pairwise equal or disjoint, and it is recovered as the quotient set of the equivalence relation it induces.

L6step 2.1step 3.1
ExampleConstruction: AI-adaptedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

A two-to-one function on a four-element set, its kernel, the quotient, and the canonical factorisation computed explicitly

Example

Write u:=u := \varnothing, v:={}v := \{\varnothing\}, w:={,{}}w := \{\varnothing,\{\varnothing\}\} and x:={{}}x := \{\{\varnothing\}\}; these four are pairwise distinct. Put A:={u,v}{w,x}A := \{u,v\} \cup \{w,x\} and B:={u,v}B := \{u,v\}, and let

f:={(u,u),(v,u)}{(w,v),(x,v)},f := \{(u,u),(v,u)\} \cup \{(w,v),(x,v)\},

so f:ABf : A \to B takes the value uu twice and the value vv twice. Then

kerf={(u,u),(u,v)}{(v,u),(v,v)}{(w,w),(w,x)}{(x,w),(x,x)},\ker f = \{(u,u),(u,v)\} \cup \{(v,u),(v,v)\} \cup \{(w,w),(w,x)\} \cup \{(x,w),(x,x)\},

A/kerf={{u,v},{w,x}},f[A]=B,A/{\ker f} = \{\{u,v\},\{w,x\}\}, \qquad f[A] = B,

and the canonical factorisation f=ιfˉπf = \iota \circ \bar f \circ \pi has π(u)=π(v)={u,v}\pi(u) = \pi(v) = \{u,v\}, π(w)=π(x)={w,x}\pi(w) = \pi(x) = \{w,x\}, fˉ({u,v})=u\bar f(\{u,v\}) = u, fˉ({w,x})=v\bar f(\{w,x\}) = v, and ι=ΔB\iota = \Delta_{B}.

Facts & Assumptions

Given: u,v,w,xu,v,w,x, AA, BB and ff as displayed.

[L3]
[L4]

[a]  :=  {bA  :  ab}    A[a] \;:=\; \{\, b \in A \;:\; a \sim b \,\} \;\subseteq\; A (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L5]

A/  :=  {[a]  :  aA}A/{\sim} \;:=\; \{\, [a] \;:\; a \in A \,\} (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L6]

bR[A]b \in R[A] holds if and only if (a,b)R(a,b) \in R for some aAa \in A (The image R[A]R[A] and the preimage R1[B]R^{-1}[B] of a set under a relation).

[L7]

{x,y}\{x,y\} is the set whose elements are exactly xx and yy, and {x}:={x,x}\{x\} := \{x,x\} (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

[L8]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

[L9]

ff is surjective (onto) if for every bBb \in B there is some xAx \in A with f(x)=bf(x) = b (Injection, surjection, bijection).

[L12]

(a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d ((a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d).

[L13]

domR:={a:b (a,b)R},ranR:={b:a (a,b)R}\operatorname{dom} R := \{\, a : \exists b\ (a,b) \in R \,\}, \qquad \operatorname{ran} R := \{\, b : \exists a\ (a,b) \in R \,\} (Relation, domR\operatorname{dom} R, ranR\operatorname{ran} R, fldR\operatorname{fld} R, and the specialisations "relation from AA to BB" and "relation on AA").

Verification

technique · direct
1.1

The four sets are pairwise distinct: uu has no element; vv and xx each have exactly one, namely uu and vv respectively, which differ; and ww has both uu and vv, so it differs from all three. Hence AA has exactly four elements and BB exactly two.

L7L8L10L11
2.1

ff is a function ABA \to B: its four pairs have distinct first coordinates, so it is single valued; its domain is AA and its range is BB. It is surjective, since u=f(u)u = f(u) and v=f(w)v = f(w), so f[A]=Bf[A] = B.

L3L6L9L12L13step 1.1
3.1

kerf\ker f is as displayed: a pair (a,a)(a,a') of elements of AA lies in it exactly when f(a)=f(a)f(a) = f(a'), and ff takes the value uu exactly at uu and vv, and the value vv exactly at ww and xx; so the pairs are those with both coordinates in {u,v}\{u,v\} or both in {w,x}\{w,x\}. By the general lemma it is an equivalence relation on AA.

L1L12step 1.1step 2.1
4.1

The classes are [u]=[v]={u,v}[u] = [v] = \{u,v\} and [w]=[x]={w,x}[w] = [x] = \{w,x\}, so A/kerfA/{\ker f} has exactly the two elements {u,v}\{u,v\} and {w,x}\{w,x\}, and the quotient map sends uu and vv to the first and ww and xx to the second.

L4L5L7step 3.1
5.1

The canonical factorisation is as displayed: fˉ\bar f sends {u,v}=[u]\{u,v\} = [u] to f(u)=uf(u) = u and {w,x}=[w]\{w,x\} = [w] to f(w)=vf(w) = v, and it is a bijection onto f[A]=Bf[A] = B; the inclusion ι\iota is the identity on BB because f[A]=Bf[A] = B; and composing gives back ff.

L2L6step 2.1step 3.1step 4.1
CounterexampleConstruction: AI-adaptedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

A symmetric and transitive relation on a two-element set that is not reflexive on it

Statement refuted

Refuted claim: a symmetric and transitive relation on a set AA is reflexive on AA. Write u:=u := \varnothing and v:={}v := \{\varnothing\}, put A:={u,v}A := \{u,v\} and

R:={(u,u)}.R := \{(u,u)\}.

RR is a relation on AA that is symmetric and transitive, and it is not reflexive on AA, because (v,v)R(v,v) \notin R.

The failure is located exactly at the point of AA that RR does not touch: RR is reflexive on its own field {u}\{u\}, and symmetry and transitivity constrain RR only there.

Facts & Assumptions

Given: u:=u := \varnothing, v:={}v := \{\varnothing\}, A:={u,v}A := \{u,v\} and R:={(u,u)}R := \{(u,u)\}.

[L1]

RR is reflexive on AA when (a,a)R(a,a) \in R for every aAa \in A (Reflexive, irreflexive, symmetric, asymmetric, antisymmetric, transitive, and connex relations on a set).

[L2]

RR is symmetric when (a,b)R(a,b) \in R implies (b,a)R(b,a) \in R, for all a,bAa, b \in A (Reflexive, irreflexive, symmetric, asymmetric, antisymmetric, transitive, and connex relations on a set).

[L3]

RR is transitive when (a,b)R(a,b) \in R and (b,c)R(b,c) \in R imply (a,c)R(a,c) \in R, for all a,b,cAa, b, c \in A (Reflexive, irreflexive, symmetric, asymmetric, antisymmetric, transitive, and connex relations on a set).

[L5]

domR:={a:b (a,b)R},ranR:={b:a (a,b)R}\operatorname{dom} R := \{\, a : \exists b\ (a,b) \in R \,\}, \qquad \operatorname{ran} R := \{\, b : \exists a\ (a,b) \in R \,\} (Relation, domR\operatorname{dom} R, ranR\operatorname{ran} R, fldR\operatorname{fld} R, and the specialisations "relation from AA to BB" and "relation on AA").

[L6]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

[L7]

{x,y}\{x,y\} is the set whose elements are exactly xx and yy, and {x}:={x,x}\{x\} := \{x,x\} (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

[L9]

(a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d ((a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d).

[L10]

(a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\} (The Kuratowski ordered pair (a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\}).

Counterexample

technique · direct
1.1

uvu \neq v, because vv has the element uu and uu has none; so AA has exactly the two elements uu and vv.

L6L7
2.1

RR is a relation on AA: its only element is the ordered pair (u,u)(u,u), and both coordinates lie in AA, so RA×AR \subseteq A \times A.

L4L8L10step 1.1
3.1

The three characteristic sets of RR are domR=ranR=fldR={u}\operatorname{dom} R = \operatorname{ran} R = \operatorname{fld} R = \{u\}, since (u,u)(u,u) is its only pair.

L5L7L9L11L12step 2.1
3.2

RR is symmetric: the only pair in RR is (u,u)(u,u), whose reversal is itself. It is transitive: the only composable pair of members is (u,u)(u,u) with (u,u)(u,u), and the required conclusion (u,u)R(u,u) \in R holds.

L2L3L9step 2.1
4.1

RR is reflexive on fldR\operatorname{fld} R: the only element of {u}\{u\} is uu, and (u,u)R(u,u) \in R.

L1L7step 3.1
5.1

RR is not reflexive on AA: vAv \in A, and (v,v)(u,u)(v,v) \neq (u,u) because vuv \neq u, so (v,v)R(v,v) \notin R. The failure is therefore confined to the single element of AA lying outside fldR\operatorname{fld} R; symmetry and transitivity say nothing about such a point, which is exactly why they do not imply reflexivity on AA.

L1L9step 1.1step 2.1step 3.1step 3.2step 4.1
False statementConstruction: Literature-sourcedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

FALSE: a symmetric and transitive relation on a set AA is reflexive on AA, so reflexivity is redundant in the definition of an equivalence relation

Statement

False statement. If a relation RR on a set AA is symmetric and transitive, then it is reflexive on AA; consequently the reflexivity clause in the definition of an equivalence relation follows from the other two and could be dropped.

Facts & Assumptions

Given: the claim above.

[L1]

RR is reflexive on AA when (a,a)R(a,a) \in R for every aAa \in A (Reflexive, irreflexive, symmetric, asymmetric, antisymmetric, transitive, and connex relations on a set).

[L2]

RR is symmetric when (a,b)R(a,b) \in R implies (b,a)R(b,a) \in R, for all a,bAa, b \in A (Reflexive, irreflexive, symmetric, asymmetric, antisymmetric, transitive, and connex relations on a set).

[L3]

RR is transitive when (a,b)R(a,b) \in R and (b,c)R(b,c) \in R imply (a,c)R(a,c) \in R, for all a,b,cAa, b, c \in A (Reflexive, irreflexive, symmetric, asymmetric, antisymmetric, transitive, and connex relations on a set).

[L4]

A binary relation \sim on AA is an equivalence relation when it is reflexive, symmetric and transitive (Equivalence relation, equivalence class, and the quotient set A/A/{\sim}).

[L7]

(a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\} (The Kuratowski ordered pair (a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\}).

[L8]

{x,y}\{x,y\} is the set whose elements are exactly xx and yy, and {x}:={x,x}\{x\} := \{x,x\} (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

[L9]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

Refutation

technique · direct
1.1

The argument that makes the claim look right: given aAa \in A, take any bb with (a,b)R(a,b) \in R; symmetry gives (b,a)R(b,a) \in R, and transitivity applied to (a,b)(a,b) and (b,a)(b,a) gives (a,a)R(a,a) \in R.

L1L2L3
1.2

The witness: put u:=u := \varnothing, v:={}v := \{\varnothing\}, A:={u,v}A := \{u,v\} and R:={(u,u)}R := \{(u,u)\}. Here uvu \neq v, because vv has an element and uu has none.

L8L9
2.1

The gap in step 1.1 is the phrase "take any bb with (a,b)R(a,b) \in R": no hypothesis supplies such a bb. Symmetry and transitivity are conditional on pairs that are already in RR, so they constrain RR only at points that RR relates to something, and say nothing whatever about a point of AA that RR leaves untouched.

L2L3step 1.1
2.2

RR is a relation on AA: its only element is the ordered pair (u,u)(u,u), whose coordinates both lie in AA.

L5L6L7step 1.2
2.3

RR is symmetric, since its only pair is its own reversal, and transitive, since the only composable pair of its members is (u,u)(u,u) with (u,u)(u,u), whose conclusion (u,u)R(u,u) \in R holds.

L2L3step 1.2
3.1

RR is not reflexive on AA: vv is an element of AA and (v,v)R(v,v) \notin R, since the only element of RR is (u,u)(u,u) and vuv \neq u.

L1L7L8step 1.2step 2.2
4.1

The claim is therefore false, and with it the conclusion drawn from it: the reflexivity clause in the definition of an equivalence relation is not redundant, since RR satisfies the other two clauses on AA and is not an equivalence relation on AA.

L4step 2.1step 2.3step 3.1
CounterexampleConstruction: AI-adaptedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

A function ff and sets S,TS, T with f[ST]f[S]f[T]f[S \cap T] \subsetneq f[S] \cap f[T]

Statement refuted

Refuted claim: f[ST]=f[S]f[T]f[S \cap T] = f[S] \cap f[T] for every function f:ABf : A \to B and all S,TAS, T \subseteq A. Write u:=u := \varnothing and v:={}v := \{\varnothing\}, and take

A:={u,v},B:={u},f:={(u,u),(v,u)},S:={u},T:={v}.A := \{u,v\}, \qquad B := \{u\}, \qquad f := \{(u,u),(v,u)\}, \qquad S := \{u\}, \qquad T := \{v\}.

Then ST=S \cap T = \varnothing, so f[ST]=f[S \cap T] = \varnothing, while f[S]=f[T]={u}f[S] = f[T] = \{u\} and therefore f[S]f[T]={u}f[S] \cap f[T] = \{u\}. The inclusion f[ST]f[S]f[T]f[S \cap T] \subseteq f[S] \cap f[T] of For f:ABf : A \to B and S,TAS, T \subseteq A: f[ST]=f[S]f[T]f[S \cup T] = f[S] \cup f[T], f[ST]f[S]f[T]f[S \cap T] \subseteq f[S] \cap f[T] and f[S]f[T]f[ST]f[S] \setminus f[T] \subseteq f[S \setminus T]; both inclusions are equalities for all SS and TT if and only if ff is injective is proper here, and it is proper exactly because ff is not injective.

Facts & Assumptions

Given: u:=u := \varnothing, v:={}v := \{\varnothing\}, A:={u,v}A := \{u,v\}, B:={u}B := \{u\}, f:={(u,u),(v,u)}f := \{(u,u),(v,u)\}, S:={u}S := \{u\} and T:={v}T := \{v\}.

[L2]

bR[A]b \in R[A] holds if and only if (a,b)R(a,b) \in R for some aAa \in A (The image R[A]R[A] and the preimage R1[B]R^{-1}[B] of a set under a relation).

[L3]

We write f:ABf : A \to B, and say ff is a function from AA to BB, when ff is a function with domf=A\operatorname{dom} f = A and ranfB\operatorname{ran} f \subseteq B (A function is a relation ff with (a,b)f(a,b) \in f and (a,c)f(a,c) \in f implying b=cb = c; f:ABf : A \to B, the value f(a)f(a), domain and codomain).

[L4]

ff is injective (one-to-one) if f(x)=f(y)f(x) = f(y) implies x=yx = y, for all x,yAx, y \in A (Injection, surjection, bijection).

[L6]

{x,y}\{x,y\} is the set whose elements are exactly xx and yy, and {x}:={x,x}\{x\} := \{x,x\} (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

[L7]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

[L8]

(a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\} (The Kuratowski ordered pair (a,b):={{a},{a,b}}(a,b) := \{\{a\},\{a,b\}\}).

[L9]

domR:={a:b (a,b)R},ranR:={b:a (a,b)R}\operatorname{dom} R := \{\, a : \exists b\ (a,b) \in R \,\}, \qquad \operatorname{ran} R := \{\, b : \exists a\ (a,b) \in R \,\} (Relation, domR\operatorname{dom} R, ranR\operatorname{ran} R, fldR\operatorname{fld} R, and the specialisations "relation from AA to BB" and "relation on AA").

Counterexample

technique · direct
1.1

uvu \neq v, since vv has an element and uu has none; so AA has exactly the two elements uu and vv, and SS and TT are the two singletons inside it.

L6L7
2.1

ff is a function ABA \to B: its two pairs have distinct first coordinates, so it is single valued; its domain is AA and its range is {u}=B\{u\} = B. It is not injective, since f(u)=u=f(v)f(u) = u = f(v) while uvu \neq v.

L3L4L8L9step 1.1
2.2

STS \cap T has no element, since an element would have to equal both uu and vv; so ST=S \cap T = \varnothing and f[ST]=f[S \cap T] = \varnothing, no element of \varnothing being related to anything.

L2L5L6L7step 1.1
3.1

f[S]={u}f[S] = \{u\} and f[T]={u}f[T] = \{u\}, since f(u)=uf(u) = u and f(v)=uf(v) = u; hence f[S]f[T]={u}f[S] \cap f[T] = \{u\}.

L2L5L6step 2.1
4.1

The general inclusion holds and is proper here: uu lies in f[S]f[T]f[S] \cap f[T] and not in f[ST]f[S \cap T], which has no element at all.

L1L7L10L11step 2.2step 3.1
ExampleConstruction: AI-generatedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5)Open item page →

For any indexed family the product iIP(Xi)\prod_{i \in I} \mathcal{P}(X_i) contains the constant function with value \varnothing, and iI{i}\prod_{i \in I} \{i\} has exactly one element

Example

Two families whose products can be shown nonempty by writing an element down, with no choice principle involved.

  • Let (Xi)iI(X_i)_{i \in I} be any indexed family. Then (P(Xi))iI(\mathcal{P}(X_i))_{i \in I} is an indexed family, and the constant function c:={(i,):iI}c := \{\,(i,\varnothing) : i \in I\,\} is an element of iIP(Xi)\prod_{i \in I} \mathcal{P}(X_i). So that product is nonempty for every II and every family, even when some XiX_i is empty.
  • Let II be any set and let ({i})iI(\{i\})_{i \in I} be the family carrying the singleton of the index at each index. Then iI{i}={ΔI}\prod_{i \in I} \{i\} = \{\Delta_I\}: its only element is the identity relation on II.

Neither construction selects anything: in the first the value is the same set at every index, and in the second the value at ii is forced to be ii.

Facts & Assumptions

Given: an indexed family (Xi)iI(X_i)_{i \in I} and a set II.

[L2]

An indexed family with index set II is a function AA with domA=I\operatorname{dom} A = I (An indexed family (Ai)iI(A_i)_{i \in I} is a function with domain II; {Ai:iI}\{A_i : i \in I\} is its range).

[L3]

zP(x)z \in \mathcal{P}(x) holds if and only if zxz \subseteq x (The power set P(x)={z:zx}\mathcal{P}(x) = \{\, z : z \subseteq x \,\}).

[L5]
[L6]

There is exactly one set with no elements, written \varnothing (There is exactly one set with no elements, written \varnothing).

[L9]

{x}:={x,x}\{x\} := \{x,x\}, the singleton of xx, is the set whose only element is xx (The unordered pair {x,y}\{x,y\} and the singleton {x}={x,x}\{x\} = \{x,x\}).

[L13]

domR:={a:b (a,b)R},ranR:={b:a (a,b)R}\operatorname{dom} R := \{\, a : \exists b\ (a,b) \in R \,\}, \qquad \operatorname{ran} R := \{\, b : \exists a\ (a,b) \in R \,\} (Relation, domR\operatorname{dom} R, ranR\operatorname{ran} R, fldR\operatorname{fld} R, and the specialisations "relation from AA to BB" and "relation on AA").

Verification

technique · direct
1.1

(P(Xi))iI(\mathcal{P}(X_i))_{i \in I} is an indexed family: separating inside I×P(P(iIXi))I \times \mathcal{P}(\mathcal{P}(\bigcup_{i \in I} X_i)) with the formula iw(z=(i,w)iIw=P(Xi))\exists i\,\exists w\,(z = (i,w) \wedge i \in I \wedge w = \mathcal{P}(X_i)) gives a set, which is single valued and has domain II; the ambient set contains each P(Xi)\mathcal{P}(X_i), since every subset of XiX_i is a subset of iIXi\bigcup_{i \in I} X_i.

L2L3L5L11L12L13L14
1.2

Similarly ({i})iI(\{i\})_{i \in I} is an indexed family, obtained by separating inside I×P(I)I \times \mathcal{P}(I), since {i}I\{i\} \subseteq I for iIi \in I.

L2L3L5L9L11L12L13
2.1

The constant function c:={(i,):iI}c := \{\,(i,\varnothing) : i \in I\,\} is a set, by separating inside I×{}I \times \{\varnothing\}; it is single valued, has domain II, and c(i)=c(i) = \varnothing for every iIi \in I. Since Xi\varnothing \subseteq X_i, we have P(Xi)\varnothing \in \mathcal{P}(X_i) for every ii, so cc lies in iIP(Xi)\prod_{i \in I} \mathcal{P}(X_i) and that product is nonempty.

L1L3L4L5L6L9L11L12L13step 1.1
2.2

ΔI\Delta_I is a function with domain II and ΔI(i)=i\Delta_I(i) = i, and ii is the only element of {i}\{i\}, so ΔIiI{i}\Delta_I \in \prod_{i \in I} \{i\}. Conversely any ff in that product has domain II and f(i){i}f(i) \in \{i\}, hence f(i)=if(i) = i for every iIi \in I, so ff and ΔI\Delta_I are functions with the same domain agreeing everywhere and are equal.

L1L5L7L8L9step 1.2
3.1

Both products are therefore nonempty, and the second has exactly one element; when I=I = \varnothing both statements agree with the general computation of the empty product, whose single element is the empty function.

L10step 2.1step 2.2
False statementConstruction: AI-adaptedVerification: AI-adaptedprecheck passverified 2026-08-06 (claude-opus-5) rests on unproved materialOpen item page →
Rests on 1 statement not proved in this library. Every dependency marked below is recorded with a citation but is not established here, because the track that would prove it has not yet been developed in this library. Everything else in this proof is proved here.

FALSE: assuming ZF is consistent, ZF proves that every surjection f:ABf : A \to B has a right inverse g:BAg : B \to A with fg=ΔBf \circ g = \Delta_B

Statement

False statement. Assume ZF is consistent. Then ZF proves that every surjection f:ABf : A \to B has a right inverse, that is, a function g:BAg : B \to A with fg=ΔBf \circ g = \Delta_B.

The consistency assumption is not decoration: an inconsistent ZF proves everything, so without it the claim would be unrefutable.

Facts & Assumptions

Given: ZF is consistent, and the claim above.

[A1]

If ZF is consistent, then ZF does not prove the Axiom of Choice (Cohen 1963, Cohen 1963: ZF does not prove the Axiom of Choice ). This is an external result, established by forcing and quoted rather than proved here.

[L1]

ff is surjective (onto) if for every bBb \in B there is some xAx \in A with f(x)=bf(x) = b (Injection, surjection, bijection).

[L2]

We write f:ABf : A \to B, and say ff is a function from AA to BB, when ff is a function with domf=A\operatorname{dom} f = A and ranfB\operatorname{ran} f \subseteq B (A function is a relation ff with (a,b)f(a,b) \in f and (a,c)f(a,c) \in f implying b=cb = c; f:ABf : A \to B, the value f(a)f(a), domain and codomain).

[L9]

An indexed family with index set II is a function AA with domA=I\operatorname{dom} A = I (An indexed family (Ai)iI(A_i)_{i \in I} is a function with domain II; {Ai:iI}\{A_i : i \in I\} is its range).

[L11]

domR:={a:b (a,b)R},ranR:={b:a (a,b)R}\operatorname{dom} R := \{\, a : \exists b\ (a,b) \in R \,\}, \qquad \operatorname{ran} R := \{\, b : \exists a\ (a,b) \in R \,\} (Relation, domR\operatorname{dom} R, ranR\operatorname{ran} R, fldR\operatorname{fld} R, and the specialisations "relation from AA to BB" and "relation on AA").

[L12]

(a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d ((a,b)=(c,d)(a,b) = (c,d) if and only if a=ca = c and b=db = d).

[L14]

An equivalent formulation of the Axiom of Choice is that a product of nonempty sets is nonempty: if XiX_i \ne \varnothing for every iIi \in I, then iIXi\prod_{i \in I} X_i \ne \varnothing (The Axiom of Choice).

Refutation

technique · contradiction
1.1

Suppose ZF proves that every surjection has a right inverse.

assume-contra
1.2

Let (Xi)iI(X_i)_{i \in I} be any indexed family with XiX_i \neq \varnothing for every iIi \in I. Separating inside I×iIXiI \times \bigcup_{i \in I} X_i gives the set E:={(i,x):iIxXi}E := \{\,(i,x) : i \in I \wedge x \in X_i\,\}, and separating inside E×IE \times I gives f:={(z,i)E×I:x(z=(i,x))}f := \{\,(z,i) \in E \times I : \exists x\,(z = (i,x))\,\}, which is a function EIE \to I sending (i,x)(i,x) to ii.

L2L8L9L10L11L12L13
2.1

ff is surjective: for iIi \in I the set XiX_i has an element xx, so (i,x)E(i,x) \in E and f((i,x))=if((i,x)) = i.

L1step 1.2
3.1

By the supposition ff has a right inverse g:IEg : I \to E with fg=ΔIf \circ g = \Delta_I. For iIi \in I we get f(g(i))=if(g(i)) = i, so g(i)=(i,x)g(i) = (i,x) for some xXix \in X_i; hence rang\operatorname{ran} g is a set of pairs, one for each iIi \in I, which is a function with domain II whose value at ii lies in XiX_i. That function is an element of iIXi\prod_{i \in I} X_i, so that product is nonempty.

L3L5L6L7L11L12step 1.1step 1.2step 2.1
4.1

So ZF would prove that a product of nonempty sets is nonempty, over an arbitrary index set, which is the product formulation of the Axiom of Choice (The Axiom of Choice); note that the hypothesis XiX_i \neq \varnothing is exactly what rules out the collapse of the product recorded in the cited computation of small products. Under the assumption that ZF is consistent, ZF does not prove the Axiom of Choice, so the supposition is untenable and the claim is false.

A1L4L14step 3.1discharge-contradiction

Remarks

Sources