Thus, the theory of nonsingular plane configurations of lines is quite different from the theory of interlacings of lines. This closely corresponds to the picture one sees in the topology of manifolds: it is known that the topology of manifolds of successive dimensions has far fewer common features than the topology of manifolds whose dimensions differ by 4.
In the topology of
high-dimensional manifolds one even has precise constructions which embed
various parts of -dimensional topology in
-dimensional topology. In
surgery theory this construction is multiplication
by a complex projective plane;
in knot theory it is the two-fold covering of Bredon; and in the theory of
singularities it is the addition to our function of the sum of the squares of
two new variables.
It seems that something similar occurs in the theory of
projective configurations. Interlacings of skew lines in three-dimensional
space appear to be related to configurations of pairwise skew
-dimensional subspaces in
-dimensional space. One can define a
linking number for oriented skew
-dimensional subspaces of
-dimensional space. Hence, all of the results on nonamphiheiral interlacings
that were proved using linking numbers carry over to this multidimensional
setting.
Moreover, there is a simple construction which to any such
configuration associates a configuration of the same type with increased by
1.
This construction preserves the linking numbers, isotopic configurations
are taken to isotopic configurations, and perhaps to some extent one has an
embedding of the theory of configurations of
-dimensional subspaces of
-dimensional space in the theory of configurations of
-dimensional subspaces of
-dimensional space. This gives rise
to the possible development of a stable theory of projective configurations.
Here we shall give a description of this construction. As far as we know,
it has not been published prior to the second version [2] of this
paper, and it was the only original result of [2].
Our construction of a suspension is applicable not only to configurations of
-dimensional subspaces of
-dimensional space.
It applies to any configuration of finitely many subspaces in projective
space; it increases the subspace dimension by 2 and the ambient space dimension
by 4.
We first recall the construction of the join of ordered configurations. Let
, ...,
be subspaces of
, and let
be subspaces of
. We imbed
and
in
as skew subspaces. In the case of odd
and
the
imbeddings should be chosen with care about orientations: the images,
with their native orientations, should have positive linking numbers in
. Let
denote the subspaces
of
such that
is the union of all lines which intersect
and
. We call the configuration of subspaces
the
join of our two configurations.
3
By the suspension of an arbitrary configuration
of
subspaces of
we mean its join with a configuration of
generatrices of a (one-sheeted) hyperboloid in
with positive
linking number (i.e., its join with the configuration of lines in
corresponding to the interlacing which we denoted by
. Since
any two lines of the interlacing
are isotropic, it follows that
one can find an isotopy of this interlacing which permutes the lines in an
arbitrary way. Hence, the join with an ordered configuration of subspaces
in
does not depend on the order. Thus, the
suspension is well defined (up to rigid isotopy) for unordered configurations.
Two configurations of -dimensional subspaces of
are
said to be stably equivalent if there exists
such that
their
-fold suspensions are rigid isotopic. Mazurovskiui[14]
has shown that this stable equivalence shares properties which are
common for various stable equivalences mentioned above. Namely,
Mazurovski
[14] has proved that for
any configuration
of
disjoint
-dimensional subspaces of
is rigidly isotopic to
the suspension of a configuration of
-dimensional subspaces of
, and, if there are
subspaces in the configurations, then
rigid isotopy of the suspensions is equivalent to rigid isotopy of the original
configurations of
-dimensional subspaces of
.
This stabilization theorem was used by Mazurovskiin [14] for
obtaining the rigid isotopy classification of nonsingular configuration
of six
-dimensional subspaces in
. He proved that
when
such configuration is defined up to rigid isotopy by the
linking numbers. Recall that this is not the case for
, that is
for configurations of lines in the 3-space. Suspension makes
configuration
shown in Figure 25 and its mirror image
rigidly isotopic. Recall that
and
are distinguished by the
Kauffman bracket polynomial. Thus, there is no generalization of
the Kauffman bracket to high-dimensional nonsingular configurations
which would be preserved under suspension.
Then Khashin and Mazurovski[15] proved that
Two nonsingular
configurations of -dimensional subspaces of
are stably
equivalent if and only if they have the same linking numbers of the
subspaces.
This means that there exists a bijection between the set of
the -subspaces of the first configuration and the set of the
-subspaces of the other configuration such that the linking numbers
of the corresponding subspaces are equal.
Algebraic techniques developed for that was used in [15] also for obtaining the following two results about interlacings of skew lines in the 3-space:
Two isotopy join interlacings are rigidly isotopic if and only if they have the same linking numbers (and hence stably equivalent).
An interlacing of skew lines which has the same linking numbers as
the configuration of disjoint generatrices of a (one-sheeted) hyperboloid in
is rigidly isotopic to this configuration of generatrices.