Math 229

Feb 27, 2009

Notes on linear algebra

Some basic ideas from linear algebra play an essential role in our subject,
so I’m o¤ering this summary as a reference. These notes won’t substitute for
a proper textbook. In particular, I won’t give proofs of anything here. Nev-
ertheless, I hope they will be of some use. I may add to these notes later the
semester. (I won’t add to them earlier).
A real vector space V is a collection of objects, referred to as vectors, that can
be added to one another and multiplied by real numbers (which are called scalars
in this context) in such a way that certain basic algebraic rules are obeyed.
The precise statement of the conditions to be met is lengthy, but contains no
surprises; roughly speaking, the rules for addition and scalar multiplication allow
us to manipulate these objects as we would expect. But for the record, here are
the rules:

1. For all v; w 2 V; v + w = w + v
2. For all u; v; w 2 V; u + (v + w) = (u + v) + w
3. There is a vector ~0 2 V such that for all v 2 V; v + ~0 = v
4. For all v 2 V; there is a vector ( v) 2 V such that v + ( v) = ~0
5. For all v; w 2 V and r 2 R; r (v + w) = rv + rw
6. For all v 2 V and r; s 2 R, (r + s) v = rv + sv
7. For all v 2 V and r; s 2 R, r (sv) = (rs) v
8. For all v 2 V; 1v = v

I’ve chosen not to put hats on all the vectors, but it’s important to remember
that there are two kinds of objects in play here, vectors and numbers. Similarly,
there are two kinds of addition going on, addition of vectors and addition of
real numbers. The same symbol “+” is used for both operations, but they are
di¤erent operations, and it would make sense (though nobody does it) to use
di¤erent symbols for them. The same comment applies to scalar multiplication
and multiplication of real numbers. I did put a hat in the symbol for the zero
vector, so that it will not be mistaken for the number 0:

Some familiar rules are not mentioned in the de…nition, like: for all v 2 V;
0v = ~0; or: for all v 2 V; ( 1) v = ( v) : These are not mentioned because they
are logical consequences of the other rules; to include them in the de…nition
would be redundant.

We can also speak about complex vector spaces, where the only di¤erence is
that we allow the scalars to be complex numbers. More generally, we can speak

1
of vector spaces where the scalars are taken from some, perhaps unspeci…ed
…eld, where the word “…eld” is a technical term that means a collection of
objects that can be added and multiplied in such a way that certain familiar
rules are satis…ed. (Not here, not now). We’ll have no need to consider …elds
other than the reals and complex numbers, but in the discussion to follow, it
doesn’t matter what the …eld is.

From here on, we assume V is a vector space. Since we don’t need to specify
the …eld of scalars associated with V , we’ll just refer to it as F .

De…nition 1 Given vectors v1 ; :::; vk in V; a linear combination of these vectors

is a vector w that is expressible in the form

X
k
w= ai vi
i=1

for some a1 ; :::; ak in F: If S is a (possibly in…nite) set of vectors in V; the span

of S is the set of all linear combinations of these vectors. If the span of S is V
itself, we say S spans V:

De…nition 2 A subspace of V is a non-empty collection W of vectors in V

that forms a vector space in its own right, using the same operations of vector
addition and scalar multiplication and the same …eld of scalars as V .

De…nition 3 A set of vectors fv1 ; :::; vk g in V is called independent providing

no one of them is a linear combination of the others.

Equivalently, a set of vectors fv1 ; :::; vk g in V is independent if no linear

combination of them is P the zero vector, unless all the chosen coe¢cients are
k
zero. That is, whenever i=1 ai vi = ~0; we must have a1 = a2 = ::: = ak = 0:
Equivalently, a set of vectors fv1 ; :::; vk g in V is independent if no two lin-
ear combinations
Pk of them
Pk are equal, unless the coe¢cients are equal. That is,
whenever i=1 ai vi = i=1 bi vi ; we must have a1 = b1 ; a2 = b2 ; :::; ak = bk :

De…nition 4 A set S of vectors in V is called a basis of V if it is independent

and it spans V:

Theorem 5 Every vector space (other than the trivial vector space consisting
of just the zero vector) has a basis. If S and S~ are bases of V; then S and S~
have the same number of elements.

Remark 6 This last assertion makes sense even in the case where S and S~ are
in…nite. In particular, not all in…nite sets have the same “number” of elements.
But all bases of a given vector space do have the same “number” of elements.

De…nition 7 The number of elements of a basis of a vector space V is called

the dimension of V; written dim (V ) :

2
Theorem 8 Every independent set in V can be extended to a basis. Every
set that spans V contains a basis. If V has dimension n (where n is a …nite
number), then every independent set consisting of n vectors is a basis. Every
spanning set consisting of n vectors is a basis.

Thus, if fv1 ; :::; vn g is a basis of V; then every vector w 2 V is expressible as

a linear combination of v1 ; :::; vn ; and in only one way. That is, for each w 2 V;
there is a unique set of scalars a1 ; a2 ; :::; an 2 F such that
X
n
w= ai vi
i=1

De…nition 9 In the above situation, the scalars a1 ; a2 ; :::; an are called the co-
ordinates of w with respect to the basis fv1 ; :::; vn g :

Now suppose that V and W are vector spaces over the same …eld F:

De…nition 10 A linear transformation T from V to W is a function from V

to W such that, for all u; v 2 V and all r; s 2 F;

T (ru + sv) = rT (u) + sT (v)

T is called an isomorphism if it is both one-to-one and onto. That is, whenever

T (u) = T (v) ; we must have u = v; and for all w 2 W; there is some v 2 V
such that T (v) = w:

Informally, if there is an isomorphism from V to W; then V and W are

indistinguishable as vector spaces. An isomorphism T from V to W carries
each vector space feature of V to a corresponding feature of W: In particular, T
must carry a basis of V to a basis of W; and (since T is one-to-one), these bases
will have the same number of elements. Therefore dim (V ) = dim (W ) : In fact
the converse is true as well. In other words, we have:

Theorem 11 Vector spaces V and W over F are isomorphic if, and only if,
they have the same dimension.

Thus, the structure of a vector space over F is completely determined by a

single number, its dimension.