Adjoint of transformation proof

zcd
Messages
197
Reaction score
0

Homework Statement


Let V be an inner product space and let T be a linear operator on V. Prove the following results:
a)R(T*)=N(T)
b)R(T*)=N(T) if V is finite dimensional

Homework Equations


<Tx,y>=<x,T*y>

The Attempt at a Solution


pick x and y ≠0 and <Tx,y>=<x,T*y>=0
this implies x∈N(T) and x ⊥ R(T*) , so N(T)⊇R(T*)

This half of the subset relationship I got, but how do I prove the other way? And as for part b), do I just prove that (W)=W? If so, why did they mention that V is finite dimensional?
 
Physics news on Phys.org
zcd said:

The Attempt at a Solution


pick x and y ≠0 and <Tx,y>=<x,T*y>=0
this implies x∈N(T) and x ⊥ R(T*) , so N(T)⊇R(T*)
i don't think this implies x∈N(T) and x ⊥ R(T*)...

but that (x∈N(T) OR Tx ⊥ y) AND x ⊥ R(T*)
 
how about starting by either assuming you have vector in the nullspace, and then for the other direction a vector perpindicular to the row vectors of T*
 
With your suggestion, I tried:

pick x and y ≠0 (the 0 case is readily proven) and x∈N(T)
then <Tx,y>=<x,T*y>=0, which implies x ⊥ R(T*) i.e x∈R(T*)
N(T)⊆R(T*)

for the other way, pick x∈R(T*) (x and y ≠0 again)
<Tx,y>=<x,T*y>=0 only when Tx=0, so x∈N(T)
N(T)⊇R(T*)

is that a sufficient proof?
 
first part pretty much, however I would write it as follows (some of it probably just diffenrent style)
-->
first pick x∈N(T)
then Tx=0
and <Tx,y>=0 for all y
but <Tx,y>=<x,T*y>=0 for all y
T*y represents a linear combination of the columns of T*
x ⊥column space of T*
x ⊥ R(T*)
so x∈R(T*)
-----

actually now that I'm working thorugh it, what exactly do you mean by R & N?
 
R is the range of transformation T and N is the nullspace of transformation T. I don't think we're supposed to use columns of T as that implies we pick a basis to represent T as a matrix.
 
ok well in the last post as y is arbitrary, you have shown x is perp to any element in the image of T (which depending on terminology can be the same as the range)
 
Back
Top