On the difference between real and complex arrangements

Let $B$ be an arrangement of linear complex hyperplanes in $C^d$. Then a classical result by Orlik \& Solomon asserts that the cohomology algebra of the complement can be constructed from the combinatorial data that are given by the intersection lattice. If $B'$ is, more generally, a $2$-arrangement in $R^{2d}$ (an arrangement of real subspaces of codimension $2$ with even-dimensional intersections), then the intersection lattice still determines the cohomology {\it groups} of the complement, as was shown by Goresky \& MacPherson. We prove, however, that for $2$-arrangements the cohomology {\it algebra} is not determined by the intersection lattice. It encodes extra information on sign patterns, which can be computed from determinants of linear relations or, equivalently, from linking coefficients in the sense of knot theory. This also allows us (in the case $d=2$) to identify arrangements with the same lattice but different fundamental groups.