Ideal property is not upper join-closed for alternating rings

From Groupprops

Statement

It is possible to have an alternating ring (i.e., an abelian group with an alternating biadditive map to itself), a subring of , and intermediate subrings of such that is an ideal in both and but not in the subring generated by and .

Related facts

Proof

Consider a four-dimensional vector space over any field, with basis . To define the product, it suffices to describe what it does on the basis elements, subject to the condition that the product of each basis element with itself is zero and the product of any two distinct basis elements is the negative of the product in the reverse order. We define the product as follows:

.

Define to be the one-dimensional subspace spanned by . Let be the subspace spanned by and and be the subspace spanned by and . Then:

  • is an ideal in : In fact, the product on is identically zero.
  • is an ideal in : In fact, the product on is identically zero.
  • is not an ideal in the subring generated by and : The subring generated by and contains . But , so is not an ideal in the subring generated.