Linear Algebra—Winter/Spring 2026
Instructors: A.S. Arora, S. Das, G. R. Kurri
Problem Set 7
(Updated: Friday, March 13, 2026)
Due: 8 PM, Tuesday March 17, 2026
Instructions
a. Timeline. In general, expect the first “half” of a problem set to be released by Tuesday
nights (or whenever the first lecture of the week takes place). Expect the full problem
set to be released by Friday nights (or whenever the second lecture of the week takes
place). The full assignment will, typically, be due on Tuesday nights, 8 PM.
b. Punctuality. Aim to get the assignment done three days prior to the deadline. We will
not be able to entertain any requests for extensions. If you seek clarification from us
and we do not respond on time, do the following: (a) assume whatever seems most
reasonable, and solve the question under that assumption and (b) ensure you have
evidence for making the clarification request and report the incident to one of the PhD
TAs.
c. Resources. Feel free to use any resource you like (including generative AI tools) to get
help but please make sure you understand what you finally write and submit. The
TAs may ask you to explain the reasoning behind your response if something appears
suspicious.
d. Graded vs Practice Questions. Problems numbered using Arabic numerals (i.e. 1, 2, 3, . . . )
constitute assignment problems that will be graded. Problems numbered using Roman
numerals (i.e. i,ii,iii,. . . ) are practice problems that will not be graded.
The text here is Linear Algebra Done Right, by Sheldon Jay Axler (3rd edition); we are on
Chapter 5 (but basic material has been taken from Chapter 3).
1. Let T be a linear operator.
a. Show that T has a 1-dimensional invariant subspace iff T has an eigenvalue.
b. Show that range(T ) is an invariant subspace of T .
c. Show that nullspace(T ) is an invariant subspace of T . Given that T has an eigen-
value λ, what can you say about nullspace(T − λI)?
2. Let V be a vector space and let S, T ∈ L(V ) be linear operators such that ST = T S.
Prove that the null space of S and the range of S are invariant under T .
3. Let P (R) denote the vector space of all real polynomials. Define the linear transfor-
mation T : P (R) → P (R) by T (p) = p′ , where p′ denotes the derivative of p. Find all
eigenvalues and the corresponding eigenvectors of T .
4. Consider the vector space V := R2 . Consider the vector v(θ) ∈ V defined as v(θ) :=
r · (cos θ, sin θ).
a. With (1, 0) along the x-axis and (0, 1) along the y-axis, schematically depict the
vector v(θ) in a figure.
b. Suppose T is a transformation that rotates all vectors in R2 (conter clock-wise)
by an angle ϕ. Write down an expression for T v(θ).
c. Show that for ϕ = π/2, a vector v = (a, b) gets mapped to T v = (−b, a).
d. Based on this geometric intuition, give an informal argument that shows that a
rotation in 2 dimensions, cannot have a real eigenvalue. Assume ϕ = π/2.
e. Prove this fact algebraically, using the operator T − λI. Assume ϕ = π/2.
5. (This asks you to essentially prove Proposition 5.26 in the text)
Suppose T ∈ L(V ) and v1 , . . . , vn is a basis of V . Show that the following are equivalent.
a. The matrix of T with respect to v1 , . . . vn is upper triangular.
b. T vj ∈ span(v1 , . . . vj ) for each j ∈ {1, . . . n}.
c. span(v1 , . . . vj ) is invariant under T for each j ∈ {1, . . . n}.
i. Let V be a vector space and let U1 , . . . , Um be subspaces of V . Suppose that
U1 + U2 + · · · + Um
is a direct sum. Prove that for every subset S ⊆ {1, 2, . . . , m}, the sum
X
Ui
i∈S
is also a direct sum.
ii. Let V be a vector space and let U1 , . . . , Um be subspaces of V . Suppose that
U1 + U2 + · · · + Um
is a direct sum. Prove that for every pair i ̸= j, Ui ∩ Uj = {0}.
Is the converse true? That is, if Ui ∩ Uj = {0} for every i ̸= j, does it follow that
U1 + U2 + · · · + Um
is a direct sum? Prove the statement or give a counterexample.
(Hint: Does something interesting happen as you go from m = 2 to m = 3).
P U . Show U1 + · · · + Um is a
iii. Suppose U1 , . . . Um are subspaces of some vector space
direct sum iff the only way of writing 0 as a sum i ui (where ui ∈ Ui ) is by taking
each uj = 0.
(It is a proposition in the text somewhere; but try thinking about it yourself; the idea
is very similar to the proof that 1:1 means dim(kernel) is zero (for linear operators)).
iv. (This is based on Example 1.43 in the text)
U1 := {(x, y, 0) : x, y ∈ F }, U2 := {(0, 0, z) : z ∈ F }, U3 := {(0, y, y) : y ∈ F }.
See if you can write (0, 0, 0) in two different ways as a sum u1 +u2 +u3 where u1 , u2 , u3 ∈
F . Also check for which i, j, it holds that Ui ∩ Uj = {0}.
v. Suppose T ∈ L(V ) is a linear operator with λ as an eigenvalue. Let U := range(T −λI).
Show that U is invariant under T .
6. [Based on Example 5.15] Suppose T is a linear operator on F 2 , defined by T (x, y) =
(0, x). Let U := {(0, y) : y ∈ F }.
a. Show that all vectors u ∈ U are eigenvectors of T with eigenvalue 0.
b. Show that the only eigenvalue T can have is 0.
c. Suppose v is an eigenvector of T . Show that v ∈ U .
d. Using your answers, show that there does not exist a subspace W that is (i)
invariant under T and (ii) is such that F 2 = U ⊕ W .
7. Suppose T ∈ L(V ) is a linear operator acting on the vector space V over the field of
complex numbers. Prove the following about T .
a. Let v ∈ V be any non-zero P vector. Show that there exist numbers a0 , a1 , . . . an ,
not all zeros, such that 0 = i∈{0,...n} ai T i v.
b. Show that a0 , a1 . . . an satisfying the equation above, are such that a0 ̸= 0.
c. Show that if c(T − λ1 I) . . . (T − λm )v = 0 then (T − λj I) is not injective, for at
least one j.