Let R = k[xi,..., xn] = (&ioRi, k an algebraically closed field of characteristic
0, and let I be a homogeneous ideal of i?, A = R/I. The Hilbert function of A,
HA : N — N, (or sometimes H(A, —)) defined by:
HA(t) = dim/e Rt - dimfe It
has been much studied. In case / is the ideal of a subscheme, X of P n _ 1 , (in which
case the Hilbert function of A = R/I is sometimes denoted Hx(—) or H(X, —))
then this function contains a great deal of information about the geometry of this
What possible functions arise in this context? This question was successfully
considered by Macaulay in .
That solution was not, however, the end of the story. Many other, related,
questions have also been considered:
i) What can H ^ be if A is a domain? (see );
ii) What can H ^ be if / = ix is the ideal of a reduced set of points, X, in
P71"1? (see );
in) What can H ^ be if / = Ix is the ideal of a set of points, X, which is the
generic hyperplane section of a curve in P
? (see , , );
iv) What can HA be if A is a Gorenstein ring? (see , , , , ,
, , , , , , , ).
We can rephrase iv) above as follows: let R be as above and let / be a homo-
geneous ideal for which \fl — (#i,... , xn). If s -f 1 is the least integer such that
C / then
4 = fc0i10...®is where As ^ 0.
The socle of A, denoted soc(A), is defined by
soc(A) : — ann^(m) where m = 0|
Since m is a homogeneous ideal of A, soc(A) is also a homogeneous ideal of A.
Clearly, As C soc(A).
soc(A) = 2li © • • • 0 2ls (noting that 2ls = As)
and let ai = dim^(2li). The integer vector
s = s(A) = (ai,...,a
is called the socle vector of A. Notice that as = dim/- 2ls ^ 0. We also call s the
socle degree of A.
It is well-known that A is a Gorenstein ring if and only if s(A) = (0,..., 0,1).