Noetherian rings and Hilbert’s basis theorem

A commutative unital ring R is Noetherian when every ideal is finitely generated, equivalently when every ascending chain of ideals stabilizes. If finite generation holds, the union of an ascending chain is an ideal generated by finitely many elements, all appearing in one chain member; no later member can be larger. Conversely, if an ideal has no finite generating set, choose a new element outside the ideal generated by the elements already selected. This constructs a strictly ascending chain, contradicting the chain condition.

The same definition applies to modules using submodules. An exact sequence has Noetherian middle module exactly when its submodule and quotient are both Noetherian. Applying this repeatedly shows finite free modules over a Noetherian ring are Noetherian, and their quotients show every finite module is Noetherian. Thus every submodule of a finite module is finite. These statements need the scalar ring’s hypothesis: an arbitrary ring can have a cyclic module whose submodule requires infinitely many generators. Noetherian is a closure property for algebraic information, not a statement about vector-space dimension.

Hilbert’s basis theorem states that R[x] is Noetherian if R is. For an ideal I of R[x], collect zero and the leading coefficients of its nonzero polynomials. They form an ideal of R: align degrees by multiplying with powers of x before adding leading terms. Choose finitely many leading coefficients generating this ideal, represented by polynomials f₁,…,f_r in I, and let d be their maximum degree. Any polynomial in I of degree at least d has its leading term canceled by a suitable R[x]-combination of these chosen polynomials.

Repeating cancellation lowers degree until a remainder has degree below d. The submodule I∩R[x]_{<d} lies in the finite free R-module of polynomials of degree below d, so it has finitely many R-module generators. Together with the f_i, they generate the whole polynomial ideal. The zero ideal can be handled separately. This proves the theorem without assuming R is a field or dividing by a leading coefficient. Iteration gives finitely many variables; infinitely many variables do not satisfy the same theorem.

Quotients and localizations of Noetherian rings are Noetherian. A finitely generated algebra over a Noetherian ring is a quotient of a finite-variable polynomial ring, so it is also Noetherian. These permanence results make the hypothesis common in algebraic geometry. It still does not force ideals to be principal: integer polynomial rings give straightforward counterexamples. Nor does it identify a canonical generating set or guarantee that a chosen redundant generating list is minimal. Those are additional structural or computational questions.

For k[x₁,x₂,…], the chain generated by the first variable, then the first two, and so on never stabilizes. Each next variable remains outside the earlier ideal, as evaluation setting the earlier variables to zero demonstrates. This counterexample explains why finite generation of an algebra matters. Also distinguish ascending ideal chains from descending ones: Z is Noetherian but its descending chain (2)⊃(4)⊃(8) continues forever. The Artinian descending-chain condition is stronger in a different direction and should not be inferred from Hilbert’s theorem.