JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

The Nullstellensatz: Sealing the Dictionary

The exact gap between I and I(V(I)) is a radical. Hilbert's Nullstellensatz says I(V(I)) = √I over an algebraically closed field, giving a perfect bijection between radical ideals and varieties — and between points and maximal ideals.

Why I and I(V(I)) differ: radicals

Consider I = (x^2) in k[x]. Then V(I) = {0}, the single point where x^2 = 0. But I({0}) = (x), strictly bigger than (x^2): the function x vanishes at the origin even though x itself was not in I. The obstruction is exactly powers: f vanishes on V(I) whenever some power f^m lies in I. The set of such f is the radical √I = { f : f^m ∈ I for some m ≥ 1 }.

An ideal is radical if √I = I. Two facts are free: every prime ideal is radical, and √I is always an ideal containing I. Geometrically V cannot see the difference between I and √I, since f and f^m have the same zero set. The natural question is whether radicals are the *only* obstruction.

The two forms of the Nullstellensatz

Hilbert's answer is yes — provided k is algebraically closed. This is the Nullstellensatz ('theorem of zero-loci'), and it comes in two faces.

Let k be ALGEBRAICALLY CLOSED, R = k[x_1, ..., x_n].

WEAK Nullstellensatz.
  If I is a proper ideal (I != R) then V(I) is NONEMPTY.
  Contrapositive: the only way to cut out the empty set is
  with the whole ring.  Equivalently, every MAXIMAL ideal
  of R has the form  m_p = (x_1 - a_1, ..., x_n - a_n)
  for a unique point  p = (a_1, ..., a_n) in A^n.

STRONG Nullstellensatz.
  For every ideal I:        I( V(I) ) = sqrt(I).

How STRONG follows from WEAK -- the 'Rabinowitsch trick':
  Suppose g vanishes on V(I), with I = (f_1,...,f_r).
  Add a new variable t and the polynomial  1 - t*g  to the
  list in R[t].  These have NO common zero (where the f_i
  vanish, g vanishes, so 1 - t*g = 1 != 0).  By WEAK, the
  ideal (f_1,...,f_r, 1 - t*g) = (1).  Write 1 as a combo,
  then substitute t = 1/g and clear denominators:
  a power  g^m  lands in I.  Hence g in sqrt(I).   QED
Weak form: empty zero set forces the unit ideal. Strong form: I(V(I)) = √I, derived by the Rabinowitsch trick.

The perfected dictionary

Now the correspondence is exact. Over an algebraically closed field, V and I are mutually inverse, order-reversing bijections between radical ideals of k[x_1,…,x_n] and affine varieties in A^n. Under it, prime ideals match irreducible varieties (guide 3) and maximal ideals match single points.