John Stillwell - Naive Lie Theory.pdf - Index of
John Stillwell - Naive Lie Theory.pdf - Index of
John Stillwell - Naive Lie Theory.pdf - Index of
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
72 3 Generalized rotation groups<br />
Maximal tori were also introduced by Weyl, in his paper Weyl [1925].<br />
In this book we use them only to find the centers <strong>of</strong> the orthogonal, unitary,<br />
and symplectic groups, since the centers turn out to be crucial in the investigation<br />
<strong>of</strong> simplicity. However, maximal tori themselves are important for<br />
many investigations in the structure <strong>of</strong> <strong>Lie</strong> groups.<br />
The existence <strong>of</strong> a nontrivial center in SO(2m), SU(n), andSp(n)<br />
shows that these groups are not simple, since the center is obviously a<br />
normal subgroup. Nevertheless, these groups are almost simple, because<br />
the center is in each case their largest normal subgroup. We have shown in<br />
Section 3.8 that the center is the largest normal subgroup that is discrete,<br />
in the sense that there is a minimum, nonzero, distance between any two<br />
<strong>of</strong> its elements. It therefore remains to show that there are no nondiscrete<br />
normal subgroups, which we do in Section 7.5.<br />
It turns out that the quotient groups <strong>of</strong> SO(2m), SU(n), andSp(n) by<br />
their centers are simple and, from the <strong>Lie</strong> theory viewpoint, taking these<br />
quotients makes very little difference. The center is essentially “invisible,”<br />
because its tangent space is zero. We explain “invisibility” in Chapter 5,<br />
after looking at the tangent spaces <strong>of</strong> some particular groups in Chapter 4.<br />
It should be mentioned, however, that the quotient <strong>of</strong> a matrix group<br />
by a normal subgroup is not necessarily a matrix group. Thus in taking<br />
quotients we may leave the world <strong>of</strong> matrix groups. The first example was<br />
discovered by Birkh<strong>of</strong>f [1936]. It is the quotient (called the Heisenberg<br />
group) <strong>of</strong> the group <strong>of</strong> upper triangular matrices <strong>of</strong> the form<br />
⎛ ⎞<br />
1 x y<br />
⎝0 1 z⎠, where x,y,z ∈ R,<br />
0 0 1<br />
by the subgroup <strong>of</strong> matrices <strong>of</strong> the form<br />
⎛<br />
1 0<br />
⎞<br />
n<br />
⎝0 1 0⎠, where n ∈ Z.<br />
0 0 1<br />
The Heisenberg group is a <strong>Lie</strong> group, but not isomorphic to a matrix group.<br />
One <strong>of</strong> the reasons for looking at tangent spaces is that we do not have<br />
to leave the world <strong>of</strong> matrices. A theorem <strong>of</strong> Ado from 1936 shows that<br />
the tangent space <strong>of</strong> any <strong>Lie</strong> group G—the <strong>Lie</strong> algebra g—can be faithfully<br />
represented by a space <strong>of</strong> matrices. And if G is almost simple then g is truly