Homework due Class 37
1. Let {v1,...,vn} be a basis of a vector space V. Let s1,...,sm be
vectors of V and we have that
s1 = a11 v1 +...+ a1n vn
s2 = a21 v1 +...+ a2n vn
........................
sm = am1 v1 +...+ amn vn
for some numbers aij. Consider the following n-vectors:
t1 = (a11, ..., a1n)
t2 = (a21, ..., a2n)
...................
tm = (am1, ..., amn)
Prove that the set {s1,...,sm} is linearly independent if and only if
the set {t1,...,tm} is linearly independent.
Let S={s1,...,sm} and T={t1,...,tm}, and let * denote scalar
multiplication.
To prove the statement "if S is linearly independent, then T is linearly
independent," assume that S is linearly independent. Let x1,...,xm be
scalars, and let the following be true:
x1*s1 + x2*s2 + ... + xm*sm = 0. (1)
Then, by the theorem about linearly independent sets of vectors, we know
that x1=x2=...=xm=0.
We may rewrite equation (1) by replacing each si with its equivalent
expression ai1*v1 + ai2*v2 + ... + a1n*vn:
x1 (a11*v1 + a12*v2 + ... + a1n*vn)
+ x2 (a21*v1 + a22*v2 + ... + a2n*vn)
+ ... + xm (am1*v1 + am2*v2 + ... + amn*vn) = 0 (2)
Then we may collect the vi terms:
(x1*a11 + x2*a21 + ... + xm*am1) v1
+ (x1*a12 + x2*a22 + ... + xm*am2) v2
+ ... + (x1*a1n + x2*a2n + ... + xm*amn) vn = 0 (3)
But {v1,...,vn} is a basis of V. So, from the definition of a basis,
{v1,...,vn} is linearly independent. So the coefficient of each vi in
equation (3) must be identically zero; that is, the following system must
be consistent:
x1*a11 + x2*a21 + ... + xm*am1 = 0
x1*a12 + x2*a22 + ... + xm*am2 = 0
................................
x1*a1n + x2*a2n + ... + xm*amn = 0.
We may rewrite this as one vector equation:
x1*(a11,...,a1n) + x2*(a21,...,a2n) + ... + xm*(am1,..,amn) = 0
and, using the definitions of t1,...,tn given above, this is equivalent
to the equation
x1*t1 + x2*t2 + ... + xm*tm = 0. (4)
By the theorem about linearly independent sets of vectors, {t1,...,tm}
will be linearly independent if the above equation has only one solution,
and if that one solution is x1=x2=...=xm=0. At the beginning of this
proof, we took a condition necessitating that, indeed,
x1=x2=...=xm=0. Since (4) is a homogeneous system of equations, it
has either exactly one solution or infinitely many solutions;
x1=x2=...=xm=0 is certainly a solution, so it is thus the only solution.
So {t1,...,tm} is linearly independent.
To prove the statement "if T is linearly independent, then S is linearly
independent," assume that T is linearly independent. Then we take
x1,...,xm to be scalars such that equation (4) above is true. So
x1=x2=...=xm=0 by the theorem about linearly independent sets of vectors.
Now we transform equation (4) back into equation (1) by proceeding
exactly backward through the steps laid out above. Equation (4) leads
directly to the equation
x1*(a11,...,a1n) + x2*(a21,...,a2n) + ...
+ xm*(am1,..,amn) = 0,
which in turn gives us the system
x1*a11 + x2*a21 + ... + xm*am1 = 0
x1*a12 + x2*a22 + ... + xm*am2 = 0
..............................
x1*a1n + x2*a2n + ... + xm*amn = 0.
This guarantees that equation (3) will be consistent in all cases, which
we may rewrite as equation (2):
x1 (a11*v1 + a12*v2 + ... + a1n*vn)
+ x2 (a21*v1 + a22*v2 + ... + a2n*vn)
+ ... + xm (am1*v1 + am2*v2 + ... + amn*vn) = 0, (2)
which is equivalent to equation (1):
x1*s1 + x2*s2 + ... + xm*sm = 0. (1)
But we know that x1=x2=...=xm=0, so (1) has only the trivial solution,
and thus the set {s1,...,sm} is linearly independent.
The proof is complete.