Lying-over theorem

In an integral extension, every prime ideal downstairs is the contraction of some prime ideal upstairs.
Lying-over theorem

Let ABA\subseteq B be an extension of . The extension is an if every element of BB is integral over AA.

Theorem (Lying over).
Assume ABA\subseteq B is an . Then for every prime ideal pA\mathfrak p\subseteq A there exists a prime ideal qB\mathfrak q\subseteq B such that

qA=p. \mathfrak q\cap A=\mathfrak p.

Equivalently, the natural map of

Spec(B)Spec(A),qqA \operatorname{Spec}(B)\longrightarrow \operatorname{Spec}(A),\qquad \mathfrak q\mapsto \mathfrak q\cap A

is surjective. In particular, every maximal ideal of AA is the contraction of some maximal ideal of BB, so the induced map on is also surjective.

Lying-over is frequently used as the existence input for and, under extra hypotheses, for .

Examples

  1. Gaussian integers over the integers.
    The inclusion ZZ[i]\mathbb Z\subset \mathbb Z[i] is integral (since ii satisfies T2+1=0T^2+1=0). For the prime ideal (5)Z(5)\subset\mathbb Z, lying-over guarantees a prime qZ[i]\mathfrak q\subset \mathbb Z[i] with qZ=(5)\mathfrak q\cap\mathbb Z=(5). Concretely, one can take q=(2+i)\mathfrak q=(2+i) (or q=(2i)\mathfrak q=(2-i)), both lying over (5)(5).

  2. A simple subring of a polynomial ring.
    Let kk be a and set A=k[t2]B=k[t]A=k[t^2]\subset B=k[t]. The element tt is integral over AA (it satisfies T2t2=0T^2-t^2=0), so ABA\subset B is integral. The prime ideal (t2)A(t^2)\subset A is the contraction of the prime ideal (t)B(t)\subset B.

  3. Adjoining a square root.
    Let A=k[x]A=k[x] and B=k[x,y]/(y2x)B=k[x,y]/(y^2-x). The image of yy is integral over AA (it satisfies T2x=0T^2-x=0), hence ABA\subset B is integral. The prime ideal (x)A(x)\subset A is the contraction of the prime ideal (x,y)B(x,y)\subset B.