Letting a, b, c, and d be the entries of A, multiplying out A<sup>2</sup> and 2A, and equating the results, I get the following cases:
. . . . .b = 0, in which case a = 2.
. . . . .b <> 0, in which case a + d = 2.
. . . . .c = 0, in which case d = 2.
. . . . .c <> 0, in which case a + d = 2.
That is, if b = 0, then a = 2. If c = 0 also, then d = 2.
If b = 0, then a = 2. If c <> 0, then, since a + d = 2 and a = 2, then d = 0.
If c = 0, then d = 2. If b = 0 also, then a = 2.
If c = 0, then d = 2. If b <> 0, then, since a + d = 2 and d = 2, then a = 0.
This gives you four matrices, fully determined. Now the messy part....
If b <> 0 and c <> 0, then all we know is that a + d = 2. But if you try just picking anything for a and d, as long as a + d = 2, you'll find that things don't work out right. So there must be something else to it.
Letting "2 - a" stand in for "d", I get:
. . . . .a<sup>2</sup> + bc = 2a
Since we know bc <> 0 (since neither b nor c is zero), then:
. . . . .a<sup>2</sup> - 2a + bc = 0
. . . . .a = [2 ± sqrt(4 - 4bc)] / 2
. . . . . . .= [2 ± 2sqrt(1 - bc)] / 2
. . . . . . .= 1 ± sqrt(1 - bc)
I don't know how helpful this is.... :shock:
Eliz.