跳转到内容
- Core constructions by their defining equivalence: z∈x∪y⇔(z∈x or z∈y); z∈x∩y⇔(z∈x and z∈y); z∈x∖y⇔(z∈x and z∈/y); y∈P(x)⇔y⊆x; y∈⋃x⇔∃z(z∈x∧y∈z).
- x=y⇔(x⊆y and y⊆x) (extensionality); pair (x,y)=(u,v)⇔x=u∧y=v; x×y={(u,v)∣u∈x,v∈y}.
- Key identities (proved by double inclusion): P(A)∩P(B)=P(A∩B); A=⋃P(A); A×⋃B=⋃{A×X∣X∈B}; monotonicity A⊆B⇒P(A)⊆P(B) and ⋃A⊆⋃B.
- Method: prove S=T via S⊆T and T⊆S; prove inclusion by taking an arbitrary element and unfolding membership.
- Relation on A = subset of A×A; aRb⇔(a,b)∈R. Properties: reflexive ∀a.aRa; symmetric; antisymmetric aRb∧bRa⇒a=b; transitive.
- Operators: R−1={(x,y)∣yRx}; S∘R={(x,z)∣∃y(xRy∧ySz)}. Facts: (S∘R)−1=R−1∘S−1; ∘ associative; R∘(S∪T)=(R∘S)∪(R∘T).
- Theorem 8 (characterizations), IA={(a,a)}: reflexive ⇔IA⊆R; symmetric ⇔R=R−1; transitive ⇔R∘R⊆R; antisymmetric ⇔R∩R−1⊆IA.
- Equivalence relation = reflexive + symmetric + transitive; class [a]R={b∣aRb}.
- aRb⇔[a]R=[b]R⇔[a]R∩[b]R=∅.
- Partition of A: nonempty, pairwise disjoint blocks with ⋃P=A. Each partition induces an equivalence relation ("same block"), whose classes are exactly the blocks.
- Transitive closure: two equivalent definitions — (universal) smallest transitive R′⊇R; (constructive) R+=⋃n∈Z+Rn with R1=R, Rn+1=Rn∘R.
- Key lemmas: Rn∘Rm=Rn+m; R+ transitive; if R⊆T and T transitive then R+⊆T (induction Rn+1=Rn∘R⊆T∘T⊆T).
- Reflexive transitive closure: R∗=⋃n∈NRn with R0=IA (= add the identity).
- F:A→B is a relation that is total (∀x∃yxFy) and single-valued; F(x) = the unique image. F=G⇔∀x.F(x)=G(x).
- (F∘G)(x)=F(G(x)). injection f(a)=f(a′)⇒a=a′; surjection ∀b∃a.f(a)=b; bijection = both.
- Composition (F:B→C, G:A→B): F,G inj ⇒F∘G inj; F,G surj ⇒F∘G surj; F∘G inj ⇒G inj; F∘G surj ⇒F surj. Bijection ⇒F−1 bijection.
- Injections f:A→B, g:B→A ⇒ a bijection A→B.
- Construction: C0=A∖ran(g), Cn+1={g(b)∣b∈Dn}, Dn=f[Cn]; set h=f on ⋃Ci, h=g−1 off it. Injectivity needs three cases (both in ⋃Ci, mixed, both off).
- A⩽B: injection exists; A≈B: bijection exists. ≈ is reflexive/symmetric/transitive.
- [0,1]≈[0,1); N≈N×N≈Q; N×N×N≈N. Products: A×B≈B×A; A1≈A2 and B1≈B2⇒A1×B1≈A2×B2.
- Function-space identities (2A=(A→{0,1}), BA=(A→B)): currying ((A×B)→C)≈(A→(B→C)); pairing (A→B×C)≈(A→B)×(A→C); ({0,1}→A)≈A×A; (A→{0,1})≈P(A) (indicator map); A≈B⇒P(A)≈P(B).
- Countable := ≈N. Countable: Z, Q, Nk, finite sequences/strings of naturals, integer polynomials, algebraic numbers.
- Cantor: X≈P(X) (diagonal Y={x∣x∈/θ(x)}). (A→{0,1})≈P(A), so P(N), NN≈2N are uncountable.
- Continuum: R≈2N≈(0,1)≈[0,1]; (R→R)≈22N; continuous R→R functions ≈R.
- Everything is a set: pair (x,y)={{x},{x,y}} (sound by pair-equality theorem); naturals 0={}, n+1=n∪{n}; N = smallest inductive set (inductive: ∅∈x, y∈x⇒y∪{y}∈x); exists and is unique.
- Addition by least T⊆N3: base (n,0,n) (n+0=n), step (n,m,r)⇒(n,m∪{m},r∪{r}) (recursion on 2nd arg). Transitive closure similarly via least T⊆A×A×N.
- Z≈(N×{0,1})∖{(0,0)}; Q≈ coprime pairs (n,m)∈Z×(N∖{0}); R = Dedekind cuts (D≈R) or Cauchy-sequence classes; C≈R×R.
- Russell's paradox: A={a∣a∈/a}⇒A∈A⇔A∈/A.
- Fix = restrict comprehension to separation {a∈A∣P(a)} (a set), while bare {a∣P(a)} is not guaranteed a set. Primitives: ∅, pairing {x,y}, union ⋃x, power set P(x). Derived as sets: x∪y=⋃{x,y}, x∩y={a∈x∣a∈y}, {x}={x,x}, x×y⊆P(P(x∪y)).