Nagata's Factoriality Theorem

• Nagata's Factoriality Theorem is a localization criterion asserting that if a noetherian integral domain’s localization at a prime-generated submonoid is a UFD, then the original ring is also a UFD.
• It relies critically on the prime-generated hypothesis, which properly accommodates products of primes rather than the overly restrictive prime-or-unit condition.
• The theorem underpins key proofs in polynomial rings and informs extensions in invariant theory and algebraic geometry by controlling divisor and primality properties.

Nagata's factoriality theorem is a localization criterion for factoriality, that is, for the unique factorization domain property. In the corrected form emphasized by recent formalization work, it states that if $R$ is a noetherian integral domain and $S \subseteq R$ is a prime-generated submonoid such that the localization $S^{-1}R$ is a UFD, then $R$ itself is a UFD (Ramos et al., 6 Apr 2026). The theorem identifies a precise descent mechanism: factoriality can pass from a localization back to the original ring when the inverted elements are finite products of primes already lying in the submonoid.

1. Statement and algebraic content

The theorem is formulated for a noetherian integral domain $R$ and a submonoid $S \subseteq R$. Its modern precise form is:

$$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$

The decisive hypothesis is that $S$ is prime-generated. This means that every element of $S$ is a finite product of elements of $S$ that are prime in $S \subseteq R$0. In the formalized presentation, this is expressed as

$S \subseteq R$1

The theorem is therefore not a statement about arbitrary localizations. It is a theorem about localizations at multiplicative systems whose elements are built from primes in a controlled way. In that sense, it is a descent theorem for factoriality rather than a general permanence theorem under localization (Ramos et al., 6 Apr 2026).

2. Why the hypotheses are structured as they are

Recent work makes explicit that the superficially simpler condition

$S \subseteq R$2

is too restrictive. If $S \subseteq R$3 are distinct nonunit primes, then $S \subseteq R$4 because $S \subseteq R$5 is a submonoid, but $S \subseteq R$6 is typically neither prime nor a unit. For that reason, the prime-or-unit condition effectively collapses to essentially one-prime situations such as

$S \subseteq R$7

The prime-generated condition is the mathematically correct replacement: it allows products of primes in $S \subseteq R$8 without requiring those products themselves to be prime (Ramos et al., 6 Apr 2026).

The noetherian hypothesis also has a specific role in the proof. It is used to place $S \subseteq R$9 in the setting of a well-founded divisibility monoid, so that every nonzero nonunit factors into irreducibles. The theorem then reduces factoriality to a primality statement for irreducibles. This shows that Nagata's theorem is not merely about localization; it depends equally on a factorization framework internal to the original ring.

A common misconception is that factoriality should descend from any UFD localization. Nagata's theorem does not assert this. The prime-generated hypothesis is the mechanism that allows divisibility data to be transferred back across localization.

3. Proof strategy and descent of primality

The proof strategy described in current formalizations is the classical one. Since $S^{-1}R$0 is noetherian, every nonzero nonunit factors into irreducibles. By the UFD criterion used there, it remains to prove that every irreducible element of $S^{-1}R$1 is prime (Ramos et al., 6 Apr 2026).

Let $S^{-1}R$2 be irreducible. The argument splits into two cases.

In the first case, $S^{-1}R$3 for some $S^{-1}R$4. Because $S^{-1}R$5 factors into primes from $S^{-1}R$6, the irreducibility of $S^{-1}R$7 forces $S^{-1}R$8 into the prime-generated structure of $S^{-1}R$9. This yields primality of $R$0 directly.

In the second case, $R$1 avoids the multiplicative system. The formalization isolates this as the predicate

$R$2

Under this avoidance hypothesis, $R$3 remains irreducible in the localization $R$4. Since $R$5 is assumed to be a UFD, $R$6 is therefore prime there. The remaining step is to descend primality from the localization back to $R$7.

The basic divisibility bridge is

$R$8

where $R$9 is the localization map. This interface allows one to translate a divisibility relation in the localization into a divisibility relation in the original ring after clearing denominators. The prime-generated structure is then used prime-by-prime to control those denominators. Thus the proof shows that an irreducible $R$0 avoiding $R$1 is prime because its image $R$2 is prime in $R$3, and that primality can be lifted back through the localization map.

4. Polynomial-ring consequences

A standard application formalized in detail is the proof that $R$4 is a UFD whenever $R$5 is a noetherian UFD. Two distinct Nagata-based proofs are packaged (Ramos et al., 6 Apr 2026).

The first proof localizes at the powers of $R$6:

$R$7

Since $R$8 is prime in $R$9, this submonoid is prime-generated. The localization $S \subseteq R$0 is identified with the Laurent polynomial ring $S \subseteq R$1, and the formalization proves that if $S \subseteq R$2 is a UFD, then $S \subseteq R$3 is a UFD. Nagata's theorem then descends factoriality from the Laurent polynomial ring back to $S \subseteq R$4.

The second proof localizes at the submonoid generated by constant primes in $S \subseteq R$5. This localization is compared with $S \subseteq R$6, which is a polynomial ring over a field and hence a UFD. Again the conclusion is obtained by Nagata descent.

The same package yields the iterated polynomial corollary

$S \subseteq R$7

A notable feature of the recent treatment is that the theorem is supplied both for the concrete localization type $S \subseteq R$8 and for abstract targets satisfying IsLocalization. This makes the theorem reusable in downstream constructions rather than tying it to a single presentation of the localization.

5. Extensions and Nagata-style analogues

Nagata's theorem has generated a broader factoriality philosophy: control of height-one primes, divisor classes, and stable localizations can force factoriality in settings far from ordinary commutative algebra.

In invariant theory, Hashimoto develops an equivariant analogue of the total ring of fractions, denoted $S \subseteq R$9, and uses it to generalize classical results of Popov on invariant and semiinvariant rings (Hashimoto, 2010). The main theorems show that, under connectedness and character-group hypotheses, the invariant ring is a UFD once the relevant $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$0-stable height-one primes contract principally. The underlying idea is explicitly Nagata-style: factoriality of invariants is governed by stable divisors and class-group data rather than by bare localization alone.

In algebraic geometry, related divisor-theoretic criteria appear in singular projective settings. For complete intersection threefolds with isolated singularities, factoriality and $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$1-factoriality coincide in the precise setting studied in "On factoriality of threefolds with isolated singularities" (Polizzi et al., 2013). In a different direction, "Factoriality of normal projective varieties" proves that for a normal projective local complete intersection with singular locus of codimension at least $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$2, $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$3-factoriality already implies factoriality, and combines this with a topological formula for the $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$4-factoriality defect to recover a projective form of Grothendieck's theorem in codimension at least $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$5 (Jung et al., 19 Jan 2026). These results are not instances of Nagata's theorem in the literal commutative-algebraic sense, but they follow the same structural principle: factoriality is detected by the disappearance of divisor-theoretic obstructions.

6. Formalization, scope, and terminological distinctions

A recent Lean 4 development provides a prime-generated formalization of Nagata's factoriality theorem and packages the result both for the concrete type Localization S and for abstract IsLocalization formulations (Ramos et al., 6 Apr 2026). The formalization is significant because it forced a correction of the hypothesis from the degenerate prime-or-unit condition to the mathematically stable prime-generated condition. It also produced reusable APIs for predicates such as PrimeGenerated and Avoids, together with transfer lemmas for divisibility, irreducibility, and primality. The authors state that no public formalization of this result is known to them in Lean, Coq, or Isabelle.

The name "Nagata's theorem" can also refer to different results. In "Nagata type statements," the central theorem attributed to Nagata asserts that if $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$6 and $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$7 are very general points in $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$8, then for $$\text{If } R \text{ is a noetherian integral domain, } S \subseteq R \text{ is prime-generated, and } S^{-1}R \text{ is a UFD, then } R \text{ is a UFD.}$$9,

$S$0

That theorem belongs to the geometry of fat points, invariant rings, Rees algebras, Mori cones, and Waldschmidt constants, and is part of the line of work surrounding Hilbert's $S$1-th problem (Roé et al., 2017). It is distinct from Nagata's factoriality theorem, although both results exemplify a broader Nagata-style method: translate an algebraic finiteness or factorization problem into structural control over prime or divisor data.

In its commutative-algebraic form, Nagata's factoriality theorem remains a canonical descent result. Its enduring content is that factoriality of a ring can be recovered from a localization, but only when the inverted multiplicative system is generated, in the precise prime-theoretic sense, by primes already visible in the original ring.