Mordell–Weil Theorem

For an abelian variety over a number field, the group of rational points is finitely generated.

Version

v2 · 2026-09-06 · History

Domain-specific #

2311

Origin domain

arithmetic geometry

Subdomain

abelian varieties

Aliases

Mordell-Weil theorem

Core Idea

The Mordell–Weil theorem states that if \(A\) is an abelian variety defined over a number field \(K\), then its group of \(K\)-rational points is finitely generated:

\[ A(K)\cong A(K)_{\mathrm{tors}}\oplus\mathbb Z^r \]

for some finite torsion subgroup and finite integer \(r\). For an elliptic curve \(E/K\), this says every rational point can be obtained by adding a finite list of generators and torsion points under the elliptic-curve group law.

The theorem is a finite-generation result, not a finiteness result: \(A(K)\) is often infinite when \(r>0\). Its autonomy lies in a reusable arithmetic-geometric bridge—algebraic points on a projective group variety become a finitely generated abelian group. Modern proofs combine a finite descent quotient with a height descent argument.^[1][2]

Structural Signature

• Number field: a finite extension \(K/\mathbb Q\).
• Abelian variety: a complete connected algebraic group \(A\) over \(K\).
• Rational-point group: \(A(K)\) is closed under the variety's group law.
• Finite quotient input: for an integer \(m\ge2\), descent establishes finiteness of \(A(K)/mA(K)\).
• Height function: arithmetic size grows quadratically under multiplication up to bounded error.
• Descent reduction: every point can be reduced modulo multiplication to a bounded-height representative.
• Northcott finiteness: only finitely many rational points have bounded degree and height.
• Finite-generation conclusion: a finite bounded set plus multiplication generates all points.
• Rank invariant: the free part has a well-defined finite rank.

Recognition test. Verify that the statement concerns rational points of an abelian variety over a number field and concludes finite generation of the group. A theorem about integral points, torsion only, or rational-point finiteness is not the Mordell–Weil theorem.

What It Is Not

It is not Mordell's conjecture/Faltings's theorem, which says a curve of genus greater than one has finitely many rational points. Abelian varieties can have infinitely many rational points while their group remains finitely generated.

It is not Siegel's theorem on integral points, not the Nagell–Lutz torsion criterion, and not Mazur's classification of rational torsion on elliptic curves over \(\mathbb Q\). Those results address different subsets or structural pieces.

It does not compute generators or rank automatically. Proving finite generation is existential. Algorithms require descent computations, height bounds, saturation, and local information. Nor does the theorem claim finite generation over arbitrary infinite algebraic extensions.

Scope of Application

For elliptic curves, the theorem gives the basic decomposition \(E(K)_{\mathrm{tors}}\oplus\mathbb Z^r\). The rank controls the infinite part and is central to Diophantine questions, descent algorithms, and the Birch–Swinnerton-Dyer conjecture. The accepted Mordell–Weil Rank node concerns this numerical invariant, not the theorem establishing its finiteness.

For higher-dimensional abelian varieties, the same finite-generation statement applies to rational points.^[3] The theorem supports arithmetic study of Jacobians: rational divisor classes on a curve map into a finitely generated group, enabling arguments about rational points on the curve.

The field hypothesis is material. Lang–Néron provides extensions over finitely generated fields with a quotient by a trace; over large infinite extensions, rational-point groups can fail to be finitely generated. The entry therefore preserves the classical number-field scope.

Clarity

Finite generation means there exist \(P_1,\ldots,P_n\in A(K)\) such that every \(P\in A(K)\) is an integer linear combination of these points. It does not mean \(A(K)\) is finite or that the representation is unique.

Because a finitely generated abelian group has a finite torsion subgroup and free part, the structural decomposition follows from the group structure theorem after finite generation is known. The theorem's difficult content is the finite-generation premise, not the abstract decomposition.

“Mordell theorem” sometimes refers specifically to elliptic curves over \(\mathbb Q\); “Weil” marks the generalization to abelian varieties over number fields.

Manages Complexity

The theorem compresses an a priori unbounded search among algebraic coordinates into finite group data: torsion, rank, and generators. Once generators are known, questions about rational-point addition reduce to integer linear combinations.

In Diophantine applications, mapping points into an abelian variety replaces geometric uncertainty with arithmetic on a finitely generated group. Heights then provide quantitative control over coefficient size.

This compression is structural, not computationally free. The theorem supplies no practical uniform bound on generators across all varieties and fields. Determining rank and proving a proposed list is complete can be difficult.

Abstract Reasoning

The standard descent lemma has the following shape. Suppose an abelian group \(G\) has a height \(h:G\to\mathbb R_{\ge0}\) such that \(h(mP)\) grows roughly as \(m^2h(P)\), height changes under translation by bounded amounts, and only finitely many points have bounded height. If \(G/mG\) is finite, choose representatives \(R_1,\ldots,R_t\).

For any sufficiently large-height \(P\), write \(P=mQ+R_i\). Height inequalities force \(h(Q)<h(P)\). Repeating eventually reaches a bounded-height point. The finite representatives and finitely many bounded points generate \(G\).

For abelian varieties over number fields, descent supplies the finite quotient and canonical heights supply the reduction mechanism. Silverman's elliptic-curve proof makes this architecture explicit.^[1]

Finite generation is qualitative rather than an automatic recipe for a basis. Determining the torsion subgroup, rank, generators, and saturation can require substantial arithmetic computation beyond the theorem's existence conclusion. Nor does a bounded-height search become finite without the Northcott-type finiteness property available in the number-field setting. These qualifications keep the statement from being misread as either an effective uniform algorithm or a theorem over arbitrary fields.

Knowledge Transfer

The proof pattern transfers within arithmetic geometry to “weak Mordell–Weil plus height descent”: first prove a quotient finite, then use a proper size function to turn finite quotient data into finite generation.

The pattern resembles generic well-founded reduction but requires arithmetic inputs: number fields, rational points, multiplication maps, local-global descent, and height finiteness. It does not license analogous claims for arbitrary groups with an informal notion of size.

Examples

1. Positive rank. An elliptic curve with \(E(\mathbb Q)\cong\mathbb Z\) has infinitely many rational points generated by one nontorsion point; this illustrates finite generation without finiteness.
2. Rank zero. If \(E(\mathbb Q)\) has rank zero, all rational points are torsion and the group is finite.
3. Higher dimension. A Jacobian \(J(C)(K)\) is finitely generated even when deciding which classes come from \(C(K)\) remains difficult.
4. Multiplication quotient. Finiteness of \(E(K)/2E(K)\) is a weak Mordell–Weil input; it is not yet the full theorem without height descent.
5. Excluded field. Passing to the algebraic closure \(\overline K\) destroys the finitely generated conclusion in general.

Structural Tensions

• Finite generation vs. finite set: the group can be infinite. Diagnostic: inspect rank rather than cardinality alone.
• Existence vs. computation: generators exist but may be hard to determine. Diagnostic: separate theorem application from completed descent and saturation.
• Weak vs. full theorem: finite \(A(K)/mA(K)\) is insufficient without height descent. Diagnostic: verify both proof ingredients.
• General variety vs. group variety: rational points on arbitrary varieties lack the abelian group law. Diagnostic: identify the abelian variety structure.
• Number field vs. large extension: the base field controls height finiteness. Diagnostic: state \(K\) and do not silently enlarge it.
• Autonomy vs. Group: Group describes the result's carrier but not arithmetic-geometric finite generation. Diagnostic: require abelian variety, number field, and height/descent conclusion.

Structural–Framed Character

The theorem is structurally a finiteness upgrade: finite quotient information plus a descending height yields finite generation. Its identity is strongly framed by arithmetic geometry. Rational points, abelian varieties, and number fields are literal.

Historical attribution is secondary to the exact statement. The name does not encompass every result called Mordell or Weil.

Structural Core vs. Domain Accent

The core is finite description of an infinite algebraic object. The domain accent is the rational-point group of an abelian variety and the descent-height proof architecture. Removing those roles leaves generic finite generation; retaining only elliptic-curve rank leaves a consequence.

Because all legitimate instances inhabit one arithmetic-geometric lineage, the theorem is domain-specific rather than prime.

Instantiates / Related Primes

Group is the proposed minimal parent by composition: the theorem's object \(A(K)\) is an abelian group, and the theorem adds a finite-generation guarantee from arithmetic geometry. Descent is proof machinery. Mordell–Weil Rank is a consequence/invariant and cannot parent the theorem.

Relationships to Other Abstractions

Current abstraction Mordell–Weil Theorem Domain-specific

Parents (1) — more general patterns this builds on

• Mordell–Weil Theorem presupposes Group Prime

  Group is the proposed minimal parent by composition: the theorem's object \(A(K)\) is an abelian group, and the theorem adds a finite-generation guarantee from arithmetic geometry.

Hierarchy paths (5) — routes to 5 parentless roots

• Mordell–Weil Theorem → Group → Monoid → Semigroup → Set and Membership

Show alternative paths (4)

Neighborhood in Abstraction Space

Mordell–Weil Theorem sits in a sparse region of the domain-specific corpus (84^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Elliptic Arithmetic & Group Finiteness (5 abstractions)

Nearest neighbors

• Néron–Tate height — 0.81
• Coin Problem — 0.81
• Schneider–Lang Theorem — 0.81
• Real Closed Field — 0.80
• Cyclic Algebra — 0.80

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

• Mordell conjecture / Faltings theorem: finiteness of rational points on high-genus curves.
• Weak Mordell–Weil theorem: finiteness of \(A(K)/mA(K)\).
• Mordell–Weil rank: rank of the free part.
• Mordell curve: a particular elliptic-curve family.
• Siegel theorem: finiteness of integral points under stated conditions.
• Lang–Néron theorem: function-field/finitely-generated-field generalization with trace qualification.

References

[1] Joseph H. Silverman, The Arithmetic of Elliptic Curves, 2^nd ed., Springer, 2009, chapters VIII–X, https://doi.org/10.1007/978-0-387-09494-6. registry ↩a ↩b

[2] Jean-Pierre Serre, Lectures on the Mordell–Weil Theorem, 3^rd ed., Vieweg, 1997, https://doi.org/10.1007/978-3-663-10632-4. registry ↩

[3] James S. Milne, Abelian Varieties, course notes, current version, sections on the Mordell–Weil theorem, https://www.jmilne.org/math/CourseNotes/AV.pdf. registry ↩