Linear-Algebra-Done-Right Β· Unit 26 Β· Video 5 Β· Interactive Practice
Why ST β TS Can Never Be the Identity
I Key Formulas
Key Insight: Every commutator S T β T S ST - TS ST β TS has trace 0 0 0 , while the identity has trace dim β‘ V \dim V dim V ; in infinite dimensions there is no trace, and the equation can hold.
II Every Commutator Has Trace 0 0 0
Whatever S S S and T T T are, the diagonal of S T β T S ST - TS ST β TS adds up to 0 0 0 , so it never matches tr β‘ I = 2 \operatorname{tr} I = 2 tr I = 2 .
III Diagonal Entries β¨ T e k , e k β© \langle Te_k, e_k\rangle β¨ T e k β , e k β β©
Turning the orthonormal basis changes each β¨ T e k , e k β© \langle Te_k, e_k\rangle β¨ T e k β , e k β β© but not their sum, tr β‘ T = 4 \operatorname{tr} T = 4 tr T = 4 for T ( x , y ) = ( x + 2 y , Β 3 y ) T(x, y) = (x + 2y,\ 3y) T ( x , y ) = ( x + 2 y , Β 3 y ) .
IV Quiz Questions
Problem 1 Β· Linearity of the Trace
Given: S , T β L ( V ) S, T \in \mathcal{L}(V) S , T β L ( V ) with tr β‘ S = 4 \operatorname{tr} S = 4 tr S = 4 and tr β‘ T = β 1 \operatorname{tr} T = -1 tr T = β 1 β find tr β‘ ( 3 S β 2 T + S T β T S ) \operatorname{tr}(3S - 2T + ST - TS) tr ( 3 S β 2 T + ST β TS ) .
a 10 10 10
b It cannot be found without knowing S S S and T T T .
c 14 14 14
d 6 6 6
β
Correct! 3 β
4 β 2 β
( β 1 ) + ( tr β‘ ( S T ) β tr β‘ ( T S ) ) = 12 + 2 + 0 = 14 3 \cdot 4 - 2 \cdot (-1) + \bigl(\operatorname{tr}(ST) - \operatorname{tr}(TS)\bigr) = 12 + 2 + 0 = 14 3 β
4 β 2 β
( β 1 ) + ( tr ( ST ) β tr ( TS ) ) = 12 + 2 + 0 = 14 .
β Check the sign. β 2 tr β‘ T = β 2 β
( β 1 ) = + 2 -2 \operatorname{tr} T = -2 \cdot (-1) = +2 β 2 tr T = β 2 β
( β 1 ) = + 2 .
β The commutator contributes nothing. tr β‘ ( S T β T S ) = tr β‘ ( S T ) β tr β‘ ( T S ) = 0 \operatorname{tr}(ST - TS) = \operatorname{tr}(ST) - \operatorname{tr}(TS) = 0 tr ( ST β TS ) = tr ( ST ) β tr ( TS ) = 0 for all
S , T S, T S , T .
β Keep the coefficients. Linearity gives
tr β‘ ( 3 S ) = 3 tr β‘ S = 12 \operatorname{tr}(3S) = 3 \operatorname{tr} S = 12 tr ( 3 S ) = 3 tr S = 12 and
tr β‘ ( β 2 T ) = 2 \operatorname{tr}(-2T) = 2 tr ( β 2 T ) = 2 .
Show solution
The trace is linear and tr β‘ ( S T ) = tr β‘ ( T S ) \operatorname{tr}(ST) = \operatorname{tr}(TS) tr ( ST ) = tr ( TS ) :
tr β‘ ( 3 S β 2 T + S T β T S ) = 3 tr β‘ S β 2 tr β‘ T + tr β‘ ( S T ) β tr β‘ ( T S ) = 12 + 2 + 0 = 14. \operatorname{tr}(3S - 2T + ST - TS) = 3\operatorname{tr} S - 2\operatorname{tr} T + \operatorname{tr}(ST) - \operatorname{tr}(TS) = 12 + 2 + 0 = 14. tr ( 3 S β 2 T + ST β TS ) = 3 tr S β 2 tr T + tr ( ST ) β tr ( TS ) = 12 + 2 + 0 = 14.
Problem 2 Β· Which Is Not a Commutator?
Given: four 2 Γ 2 2 \times 2 2 Γ 2 matrices β find the one that is NOT of the form M ( S T β T S ) \mathcal{M}(ST - TS) M ( ST β TS ) for any S , T β L ( C 2 ) S, T \in \mathcal{L}(\mathbf{C}^2) S , T β L ( C 2 ) (matrices with respect to the standard basis).
a ( 0 1 0 0 ) \begin{pmatrix} 0 & 1 \\ 0 & 0 \end{pmatrix} ( 0 0 β 1 0 β )
b ( 1 0 0 β 1 ) \begin{pmatrix} 1 & 0 \\ 0 & -1 \end{pmatrix} ( 1 0 β 0 β 1 β )
c ( 0 0 0 0 ) \begin{pmatrix} 0 & 0 \\ 0 & 0 \end{pmatrix} ( 0 0 β 0 0 β )
d ( 1 0 0 0 ) \begin{pmatrix} 1 & 0 \\ 0 & 0 \end{pmatrix} ( 1 0 β 0 0 β )
β
Correct! Its trace is
1 1 1 , but every
S T β T S ST - TS ST β TS has trace
0 0 0 .
β That one is a commutator. With
S = ( 1 0 0 0 ) S = \begin{pmatrix} 1 & 0 \\ 0 & 0 \end{pmatrix} S = ( 1 0 β 0 0 β ) and
T = ( 0 1 0 0 ) T = \begin{pmatrix} 0 & 1 \\ 0 & 0 \end{pmatrix} T = ( 0 0 β 1 0 β ) ,
S T β T S = T β 0 = T ST - TS = T - 0 = T ST β TS = T β 0 = T .
β That one is a commutator. With
S = ( 0 1 0 0 ) S = \begin{pmatrix} 0 & 1 \\ 0 & 0 \end{pmatrix} S = ( 0 0 β 1 0 β ) and
T = ( 0 0 1 0 ) T = \begin{pmatrix} 0 & 0 \\ 1 & 0 \end{pmatrix} T = ( 0 1 β 0 0 β ) ,
S T β T S = ( 1 0 0 0 ) β ( 0 0 0 1 ) ST - TS = \begin{pmatrix} 1 & 0 \\ 0 & 0 \end{pmatrix} - \begin{pmatrix} 0 & 0 \\ 0 & 1 \end{pmatrix} ST β TS = ( 1 0 β 0 0 β ) β ( 0 0 β 0 1 β ) .
β That one is a commutator. Take
S = T S = T S = T : then
S T β T S = 0 ST - TS = 0 ST β TS = 0 .
Show solution
Since tr β‘ ( S T β T S ) = tr β‘ ( S T ) β tr β‘ ( T S ) = 0 \operatorname{tr}(ST - TS) = \operatorname{tr}(ST) - \operatorname{tr}(TS) = 0 tr ( ST β TS ) = tr ( ST ) β tr ( TS ) = 0 , a matrix with nonzero trace is never a commutator. Option (d) has trace 1 1 1 .
The other three have trace 0 0 0 and are commutators: (a) = [ 1 0 0 0 ] [ 0 1 0 0 ] β [ 0 1 0 0 ] [ 1 0 0 0 ] = \bigl[\begin{smallmatrix} 1 & 0 \\ 0 & 0 \end{smallmatrix}\bigr]\bigl[\begin{smallmatrix} 0 & 1 \\ 0 & 0 \end{smallmatrix}\bigr] - \bigl[\begin{smallmatrix} 0 & 1 \\ 0 & 0 \end{smallmatrix}\bigr]\bigl[\begin{smallmatrix} 1 & 0 \\ 0 & 0 \end{smallmatrix}\bigr] = [ 1 0 β 0 0 β ] [ 0 0 β 1 0 β ] β [ 0 0 β 1 0 β ] [ 1 0 β 0 0 β ] ; (b) = [ 0 1 0 0 ] [ 0 0 1 0 ] β [ 0 0 1 0 ] [ 0 1 0 0 ] = \bigl[\begin{smallmatrix} 0 & 1 \\ 0 & 0 \end{smallmatrix}\bigr]\bigl[\begin{smallmatrix} 0 & 0 \\ 1 & 0 \end{smallmatrix}\bigr] - \bigl[\begin{smallmatrix} 0 & 0 \\ 1 & 0 \end{smallmatrix}\bigr]\bigl[\begin{smallmatrix} 0 & 1 \\ 0 & 0 \end{smallmatrix}\bigr] = [ 0 0 β 1 0 β ] [ 0 1 β 0 0 β ] β [ 0 1 β 0 0 β ] [ 0 0 β 1 0 β ] ; (c) = S S β S S = SS - SS = SS β SS for any S S S .
Problem 3 Β· Diagonal Entries in an Orthonormal Basis
Given: T β L ( R 2 ) T \in \mathcal{L}(\mathbf{R}^2) T β L ( R 2 ) defined by T ( x , y ) = ( 2 x + y , Β β x + 4 y ) T(x, y) = (2x + y,\ -x + 4y) T ( x , y ) = ( 2 x + y , Β β x + 4 y ) , the Euclidean inner product, and the orthonormal basis e 1 = 1 5 ( 3 , 4 ) e_1 = \tfrac15(3, 4) e 1 β = 5 1 β ( 3 , 4 ) , e 2 = 1 5 ( β 4 , 3 ) e_2 = \tfrac15(-4, 3) e 2 β = 5 1 β ( β 4 , 3 ) β find β¨ T e 1 , e 1 β© \langle Te_1, e_1\rangle β¨ T e 1 β , e 1 β β© and β¨ T e 2 , e 2 β© \langle Te_2, e_2\rangle β¨ T e 2 β , e 2 β β© .
β¨ T e 1 , e 1 β© \langle Te_1, e_1\rangle β¨ T e 1 β , e 1 β β©
a 82 82 82
b 82 25 \tfrac{82}{25} 25 82 β
c 2 2 2
d 82 5 \tfrac{82}{5} 5 82 β
β¨ T e 2 , e 2 β© \langle Te_2, e_2\rangle β¨ T e 2 β , e 2 β β©
a 4 4 4
b 82 25 \tfrac{82}{25} 25 82 β
c 68 25 \tfrac{68}{25} 25 68 β
d β 68 25 -\tfrac{68}{25} β 25 68 β
β
Correct! 82 25 + 68 25 = 6 = 2 + 4 = tr β‘ T \tfrac{82}{25} + \tfrac{68}{25} = 6 = 2 + 4 = \operatorname{tr} T 25 82 β + 25 68 β = 6 = 2 + 4 = tr T .
β e 1 e_1 e 1 β has length 1 1 1 only with the factor 1 5 \tfrac15 5 1 β . 82 = β¨ T ( 3 , 4 ) , ( 3 , 4 ) β© 82 = \langle T(3, 4), (3, 4)\rangle 82 = β¨ T ( 3 , 4 ) , ( 3 , 4 )β© must be divided by
5 2 5^2 5 2 .
β 2 2 2 is the entry for the standard basis vector ( 1 , 0 ) (1, 0) ( 1 , 0 ) . Diagonal entries depend on the orthonormal basis; only their sum does not.
β 4 4 4 is the entry for ( 0 , 1 ) (0, 1) ( 0 , 1 ) . Compute
β¨ T e 2 , e 2 β© \langle Te_2, e_2\rangle β¨ T e 2 β , e 2 β β© for
e 2 = 1 5 ( β 4 , 3 ) e_2 = \tfrac15(-4, 3) e 2 β = 5 1 β ( β 4 , 3 ) , or subtract from
tr β‘ T = 6 \operatorname{tr} T = 6 tr T = 6 .
β The two entries need not be equal. They must add up to
tr β‘ T = 6 \operatorname{tr} T = 6 tr T = 6 .
β Not quite. T e 1 = 1 5 ( 10 , 13 ) Te_1 = \tfrac15(10, 13) T e 1 β = 5 1 β ( 10 , 13 ) ; take its inner product with
1 5 ( 3 , 4 ) \tfrac15(3, 4) 5 1 β ( 3 , 4 ) .
β Not quite. The two diagonal entries add up to
tr β‘ T = 2 + 4 = 6 \operatorname{tr} T = 2 + 4 = 6 tr T = 2 + 4 = 6 .
Show solution
T e 1 = 1 5 ( 6 + 4 , Β β 3 + 16 ) = 1 5 ( 10 , 13 ) , β¨ T e 1 , e 1 β© = 1 25 ( 30 + 52 ) = 82 25 . Te_1 = \tfrac15(6 + 4,\ -3 + 16) = \tfrac15(10, 13), \qquad \langle Te_1, e_1\rangle = \tfrac1{25}(30 + 52) = \tfrac{82}{25}. T e 1 β = 5 1 β ( 6 + 4 , Β β 3 + 16 ) = 5 1 β ( 10 , 13 ) , β¨ T e 1 β , e 1 β β© = 25 1 β ( 30 + 52 ) = 25 82 β .
T e 2 = 1 5 ( β 8 + 3 , Β 4 + 12 ) = 1 5 ( β 5 , 16 ) , β¨ T e 2 , e 2 β© = 1 25 ( 20 + 48 ) = 68 25 . Te_2 = \tfrac15(-8 + 3,\ 4 + 12) = \tfrac15(-5, 16), \qquad \langle Te_2, e_2\rangle = \tfrac1{25}(20 + 48) = \tfrac{68}{25}. T e 2 β = 5 1 β ( β 8 + 3 , Β 4 + 12 ) = 5 1 β ( β 5 , 16 ) , β¨ T e 2 β , e 2 β β© = 25 1 β ( 20 + 48 ) = 25 68 β .
The sum is 150 25 = 6 \tfrac{150}{25} = 6 25 150 β = 6 , which is tr β‘ T = 2 + 4 \operatorname{tr} T = 2 + 4 tr T = 2 + 4 computed from the standard basis.
Problem 4 Β· A Scaled Identity as a Commutator
Given: V V V with dim β‘ V = 5 \dim V = 5 dim V = 5 , and Ξ» β F \lambda \in \mathbf{F} Ξ» β F such that S T β T S = Ξ» I ST - TS = \lambda I ST β TS = Ξ» I for some S , T β L ( V ) S, T \in \mathcal{L}(V) S , T β L ( V ) β find which values of Ξ» \lambda Ξ» are possible.
a only Ξ» = 0 \lambda = 0 Ξ» = 0
b every Ξ» β F \lambda \in \mathbf{F} Ξ» β F
c only Ξ» = 5 \lambda = 5 Ξ» = 5
d no value of Ξ» \lambda Ξ»
β
Correct! Taking traces gives
0 = 5 Ξ» 0 = 5\lambda 0 = 5 Ξ» , so
Ξ» = 0 \lambda = 0 Ξ» = 0 , and
Ξ» = 0 \lambda = 0 Ξ» = 0 does occur (take
S = T S = T S = T ).
β Take the trace of both sides. tr β‘ ( S T β T S ) = 0 \operatorname{tr}(ST - TS) = 0 tr ( ST β TS ) = 0 while
tr β‘ ( Ξ» I ) = 5 Ξ» \operatorname{tr}(\lambda I) = 5\lambda tr ( Ξ» I ) = 5 Ξ» .
β 5 5 5 is tr β‘ I \operatorname{tr} I tr I , not a possible Ξ» \lambda Ξ» . The trace equation is
0 = 5 Ξ» 0 = 5\lambda 0 = 5 Ξ» .
β Ξ» = 0 \lambda = 0 Ξ» = 0 is possible. Any operator commutes with itself, so
S S β S S = 0 = 0 β
I SS - SS = 0 = 0 \cdot I SS β SS = 0 = 0 β
I .
Show solution
Apply the trace, which is linear and satisfies tr β‘ ( S T ) = tr β‘ ( T S ) \operatorname{tr}(ST) = \operatorname{tr}(TS) tr ( ST ) = tr ( TS ) :
0 = tr β‘ ( S T ) β tr β‘ ( T S ) = tr β‘ ( S T β T S ) = tr β‘ ( Ξ» I ) = Ξ» dim β‘ V = 5 Ξ» . 0 = \operatorname{tr}(ST) - \operatorname{tr}(TS) = \operatorname{tr}(ST - TS) = \operatorname{tr}(\lambda I) = \lambda \dim V = 5\lambda. 0 = tr ( ST ) β tr ( TS ) = tr ( ST β TS ) = tr ( Ξ» I ) = Ξ» dim V = 5 Ξ» .
So Ξ» = 0 \lambda = 0 Ξ» = 0 . Conversely Ξ» = 0 \lambda = 0 Ξ» = 0 is attained, for instance by S = T = I S = T = I S = T = I . The case Ξ» = 1 \lambda = 1 Ξ» = 1 is the video's statement that S T β T S = I ST - TS = I ST β TS = I is impossible.
Solved: 0 / 4