Questions: Successor Ordinals, Limit Ordinals, and Von Neumann Construction

Question 1

According to the Von Neumann construction, what is the ordinal 3?

• A. The number that comes after 2, with no further set-theoretic definition
• B. {∅, {∅}, {∅, {∅}}} — the set containing 0, 1, and 2
• C. {{∅}} — the set containing the set containing the empty set
• D. The successor of 2, defined as S(2) = {2} alone

Answer: (B)

In the Von Neumann construction, each ordinal is the set of all ordinals that came before it: 0 = ∅, 1 = {0} = {∅}, 2 = {0,1} = {∅,{∅}}, 3 = {0,1,2} = {∅,{∅},{∅,{∅}}}. The successor operation adds the ordinal itself as a new element: S(α) = α ∪ {α}. So 3 = S(2) = 2 ∪ {2} = {0,1} ∪ {2} = {0,1,2}. Every ordinal contains exactly its predecessors as elements — a feature that makes the ordering relation (α < β iff α ∈ β) coincide perfectly with set membership.

Question 2

Why does transfinite induction require three cases rather than the two cases (base + inductive step) of ordinary mathematical induction?

• A. Because ordinals extend into three 'zones': finite, countably infinite, and uncountably infinite
• B. Because limit ordinals like ω have no immediate predecessor, so the successor step cannot reach them from below
• C. Because the base case at 0 must be split into two sub-cases for even and odd ordinals
• D. Because set theory requires an extra case to handle the axiom of choice

Answer: (B)

Ordinary induction proves P(0) and (P(n) → P(n+1)), which covers all natural numbers because every natural number can be reached by repeatedly applying the successor from 0. But ω is a limit ordinal — there is no ordinal just below it, so no single successor step reaches it. Transfinite induction adds a limit case: if P(α) holds for all α < λ (for every limit ordinal λ), then P(λ) holds. This third case is necessary precisely because limit ordinals are defined as suprema, not as successors, and the proof must reflect that structure.

Question 3

In the Von Neumann construction, the statement 'α < β' (α is less than β as ordinals) is equivalent to 'α ∈ β' (α is a member of β as sets).

• T. True
• F. False

Answer: True

This is the elegant feature of the Von Neumann construction: the two most natural relations on ordinals (order and membership) coincide exactly. Since each ordinal is defined as the set of all smaller ordinals, β = {α : α < β} — so α is an element of β if and only if α is less than β. This makes ordinal comparisons set-theoretically transparent: to check if one ordinal is smaller than another, you check membership. No separate definition of '<' is needed; it is inherited from ∈.

Question 4

ω (the first infinite ordinal) is a successor ordinal — it is the successor of the largest finite ordinal.

• T. True
• F. False

Answer: False

ω is a limit ordinal, not a successor ordinal. There is no largest finite ordinal: for every finite ordinal n, S(n) = n+1 is also a finite ordinal. Since no finite ordinal is 'just below' ω, ω cannot be reached by a single successor step. Instead, ω is defined as the set of all finite ordinals: ω = {0, 1, 2, 3, …} — the supremum of the infinite sequence of finite ordinals. This is the defining characteristic of a limit ordinal: it has no immediate predecessor and equals the union (supremum) of all ordinals below it.

Question 5

What does it mean to say that ω is a 'limit ordinal,' and why does this make it fundamentally different from finite ordinals?

Model answer: ω is a limit ordinal because it has no immediate predecessor — there is no ordinal α such that S(α) = ω. Every finite ordinal n is a successor ordinal: n = S(n−1). But ω is the supremum of the entire infinite sequence {0, 1, 2, 3, …}; you cannot reach it by any finite number of successor steps from below. In the Von Neumann construction, ω is defined as the set of all finite ordinals: ω = {0, 1, 2, …}, so it contains infinitely many elements. This is why transfinite induction needs a special limit case: you cannot use the successor step to 'get to' ω, so you must prove the property holds at ω by a different argument — typically showing it holds for all smaller ordinals and that this forces it to hold at the supremum.

The limit/successor distinction is fundamental to the structure of transfinite arithmetic. It recurs at every level: after ω come ω+1, ω+2, … (successor ordinals), then the limit ordinal ω·2, and so on. The ordinals alternate between successor stages (reachable by one step) and limit stages (defined as suprema). Any proof or recursive definition over all ordinals must handle both types explicitly.