Questions: Forking and Independence in Stability Theory

Question 1

In a stable theory, what does it mean for a type p(x) over B to fork over A ⊆ B?

• A. p has more realizations in the model than expected given A
• B. p implies a finite disjunction of formulas, each with only finitely many realizations, using parameters from B that are not in A
• C. p is inconsistent with the complete theory when parameters from B are added
• D. p has no non-forking extensions to any set larger than B

Answer: (B)

Forking captures the idea that B contains new information that dramatically constrains where x can land — information absent from A. Technically, p forks over A if it implies φ₁(x,b₁) ∨ ... ∨ φₙ(x,bₙ) where each φᵢ has only finitely many realizations (is 'algebraically constraining'). A forking type is still consistent — it just has become dependent on the new parameters. Options A and C describe different phenomena entirely.

Question 2

In an algebraically closed field (ACF), you are given a type of a single element a over a set A. Under what condition does this type fork over A?

• A. When a is transcendental over A — that is, a satisfies no polynomial over A
• B. When a is algebraically dependent over A — it satisfies a polynomial equation with coefficients in A
• C. When a is contained in the algebraic closure of A but not literally in A itself
• D. When A is uncountable, causing the type space to be too large

Answer: (B)

In ACF, non-forking corresponds exactly to algebraic independence. Forking therefore corresponds to algebraic dependence: a forks over A when it satisfies a polynomial with coefficients in A, meaning B has 'algebraically constrained' a beyond what A already determined. Option A describes the non-forking case — transcendental elements are the independent ones. Option C is incorrect because an element in the algebraic closure (but not A) can still be in a forking or non-forking extension depending on the type.

Question 3

Forking independence in stable theories is symmetric: if a is independent from b over A, then b is independent from a over A.

• T. True
• F. False

Answer: True

Symmetry is one of the four key axioms that forking satisfies in stable theories. This is genuinely non-obvious — 'a is independent from b over A' and 'b is independent from a over A' are logically different statements about different types, yet in stable theories they are equivalent. Symmetry is what makes forking a true independence relation rather than a one-directional dependence notion.

Question 4

In algebraically closed fields, forking independence is determined solely by the cardinality of the tuple: any two tuples of the same length over A either both fork or both do not fork over A.

• T. True
• F. False

Answer: False

Cardinality is irrelevant to whether a tuple forks over A. Two tuples of the same length can behave completely differently: one may be algebraically independent over A (non-forking) while the other satisfies polynomial relations over A (forking). Forking depends on the specific algebraic relationships between the elements and A, not on how many elements there are.

Question 5

What role does forking play in stable theories that is analogous to the role linear independence plays in vector spaces?

Model answer: Forking defines a dimension theory for stable structures. Just as linear independence measures how many vectors contribute genuinely new 'directions' and yields the dimension of a vector space, non-forking independence measures how many elements are 'genuinely new' over a base set and yields a dimension (Morley rank) for stable structures. The forking geometry of a strongly minimal set then classifies the structure geometrically — modular (vector-space-like), projective, or more complex.

The analogy is deep: both linear independence and forking independence satisfy the exchange property (in appropriate forms), which is what makes a dimension theory possible. In ACF, non-forking corresponds to algebraic independence and the dimension is transcendence degree. The classification of stable theories by their forking geometry — one of the deepest results of model theory — rests entirely on this analogy.