Problem 1. Construct (with proof) bases for the following vector spaces.
(a) The subspace of R3 defined as the set of vectors (x, y, z) R3 satisfying the conditions x + y + z =
x y 2z = 0.
(b) The subspace of P4 (R) consisting of polynomials f (x) satisfying x4 f (1/x) = f (x).
(c) The quotient vector space V /W over R, where V = R2 and W is the 1-dimensional subspace determined
by the equation x = y.
Problem 2. Let V be a finite-dimensional vector space over a field F and W a subspace. Use the basis
extension theorem to prove that there exists a subspace W 0 such that W + W 0 = V and W W 0 = 0.
Problem 3. If F is a field, we know that the space Pn (F ) of polynomials of degree n has dimension
n + 1, by taking the standard basis 1, x, . . . , xn . This problem discusses how to construct a different basis
of Lagrange interpolation polynomials.
(a) Let a0 , . . . , an be n + 1 distinct elements of F . Define polynomials 0 , . . . , n by
Y (x aj )
i (x) = ,
(ai aj )
j6=i
where this notation means the product over all indices 0 j n, excluding the index j = i. Prove that
these n + 1 polynomials are linearly independent and thus are a basis for Pn (F ). (Hint: Look at what
happens when you evaluate the polynomials at the n + 1 points a0 , . . . , an ).
(b) Write out the basis explicitly for P3 (R) with {a0 , . . . , a3 } = {0, 1, 2, 3}.
(c) Rephrasing part (a) we can say that if we choose n + 1 distinct points, we can construct a basis of Pn (F )
by specifying the values of polynomials at those n + 1 points. For F = R, say, we can use this idea to produce
many bases for Pn (R) for any n (by choosing any list of n + 1 points). But if we consider the finite field
F = Fp = Z/pZ, we can only get bases of Pn (Fp ) this way for n < p. It turns out that over finite fields we
have the strange phenomenon that a nonzero polynomial can take the value zero everywhere! Demonstrate
this by showing that if we consider the polynomial f (x) = xp x Pp (Fp ), then all of the values f (a) for
a Fp are zero. (You can use the result called Fermats Little Theorem without needing to prove it.)
Problem 4. Let V be a finite-dimensional vector space over a field F , and let S, T, U be subspaces.
(a) If S T = 0, prove that dim(S + T ) = dim S + dim T .
(b) In general, prove dim(S + T ) = dim S + dim T dim(S T ).
(c) If youre familiar with the inclusion-exclusion principle, you might guess that part (b) generalizes to
the equality
Provide a counterexample to show this equality is false in general! (Hint: You can take the field F to be
R and the vector space V to be R2 ).
Extended Glossary. Give a definition of an internal direct sum decomposition of a vector space.
Give an example and a non-example of such a decomposition. Then state and prove a theorem about internal
direct sums.