Invertible linear transformation

Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
6 replies · 2K views
yifli
Messages
68
Reaction score
0

Homework Statement


State why it is that if V is finite-dimensional, and if S and T in HomV satisfy [tex]S\circ T = I[/tex], then T is invertible and S=T^-1



2. The attempt at a solution
Isn't this obvious? I don't understand why any elaboration is necessary
 
Physics news on Phys.org
Hi yifli! :smile:

No, this isn't obvious. T invertible means

[tex]S\circ T=1~\text{and}~T\circ S=1[/tex].

Now you only have that [itex]S\circ T=1[/itex]. Somehow, you need to show [itex]T\circ S=1[/itex].

(Hint: use determinants)
 
You don't need to use determinants (or even matrices), and in fact doing so might introduce circular logic depending on what theorems you use.

Instead, answer this:

[tex]S\circ T = I[/tex]

implies that T is _____ and S is _____

(Fill in the blanks with either injective or surjective as appropriate.)

Then you'll need to use the finite dimensionality of V to draw a conclusion.
 
jbunniii said:
You don't need to use determinants (or even matrices), and in fact doing so might introduce circular logic depending on what theorems you use.

Instead, answer this:

[tex]S\circ T = I[/tex]

implies that T is _____ and S is _____

(Fill in the blanks with either injective or surjective as appropriate.)

Then you'll need to use the finite dimensionality of V to draw a conclusion.

implies T is injective and S is surjective
since V is finite dimensional, T is an isomorphism,so S = T^-1
 
micromass said:
Hi yifli! :smile:

No, this isn't obvious. T invertible means

[tex]S\circ T=1~\text{and}~T\circ S=1[/tex].

Now you only have that [itex]S\circ T=1[/itex]. Somehow, you need to show [itex]T\circ S=1[/itex].

(Hint: use determinants)

Can you give me an example that shows T is not invertible if only ST=I is given?

thanks
 
yifli said:
Can you give me an example that shows T is not invertible if only ST=I is given?

thanks

It has to be an infinite-dimensional example, as you just showed.

Let V be the vector space consisting of infinite sequences of the form

[tex](a_1, a_2, a_3, \ldots)[/tex]

where the [tex]a_i[/tex] are real numbers. (Or complex. It doesn't matter.)

Addition and scalar multiplication are defined pointwise in the obvious way.

Let T be the operator that shifts the sequence right by one step:

[tex]T(a_1, a_2, a_3, \ldots) = (0, a_1, a_2, \ldots)[/tex]

And let S be the operator that shifts the sequence left by one step:

[tex]S(a_1, a_2, a_3, \ldots) = (a_2, a_3, a_4, \ldots)[/tex]

You can verify that these are linear operators, and that

[tex]S \circ T = I[/tex]

but

[tex]T \circ S \neq I[/tex]

T can be "undone" but S cannot.
 
yifli said:
Can you give me an example that shows T is not invertible if only ST=I is given?

thanks

With other things than vector spaces you mean?
Well, take T:{1}-> {1,2} with T(1)=1 and take S:{1,2}->{1} with S(1)=1 and S(2)=1.