The University of Sydney

MOOC Introduction to Linear Algebra

Notes to accompany ‘Dot Product of Two Vectors’

Key ideas and useful facts:

(i) Geometric definition of dot product of two vectors: If v and w are geometric vectors then
their dot product v · w can be defined geometrically as follows:
v · w = |v||w| cos θ
where θ is the angle between them.

This formula is related to the projection of the tip of one vector onto the line passing
through the other vector (assuming the vectors emanate from the same point). For
example, if we project the tip of v onto the line passing through w we get the following:

w
ℓ
If we denote the length of this projection by ℓ then
v·w
ℓ = |v| cos θ = .
|w|
(This assumes that θ is acute. If the angle is obtuse then we need to add a negative sign.)

1
 (ii) Algebraic definition of dot product of two vectors: If v and w are geometric vectors with
the following Cartesian decompositions:

v = ai + bj + ck and w = di + ej + f k ,

then their dot product v · w can be defined algebraically as follows:

v · w = ad + be + cf ,

that is, multiplying the respective components together and then taking the sum. It is
explained below why the geometric and algebraic definitions in fact coincide, even though
superficially they appear to be completely different.
(iii) Length of a vector in terms of the dot product: If v is a geometric vector with Cartesian
decomposition v = ai + bj + ck then

v · v = a2 + b2 + c2 = |v|2 ,

so that √ √
|v| = a2 + b2 + c2 = v·v .
√
(The conclusion |v| = v · v also follows immediately from the geometric definition of
the dot product, using the fact that the angle between v and itself is zero and cos 0 = 1.)
(iv) Angle between vectors in terms of the dot product: If one knows in advance the value of the
dot product v · w of geometric vectors v and w, then one can find the angle θ between
them using the following formula:
v·w
cos θ = .
|v||w|
If the dot product is a positive real number then θ is acute or zero. If the dot product
is a negative real number then θ is obtuse or straight (180◦ ). If the dot product is zero
then θ is a right angle, in which case v and w are mutually perpendicular.
(v) Cauchy-Schwarz inequality: If v and w are geometric vectors then

|v · w| ≤ |w||v| .

This follows quickly from the geometric definition of the dot product because | cos θ| ≤ 1
always. To deduce this inequality from the algebraic definition of the dot product is quite
difficult.
(vi) Commutativity: The dot product is commutative in the sense that if v and w are geometric
vectors then
v·w = w·v .
This property follows immediately from either of the geometric or algebraic definitions of
the dot product.
(vii) Distributivity: The dot product is distributive over addition in the sense that if u, v and
w are geometric vectors then

(u + v) · w = u · w + v · w .

We carefully verify this property below.

2
(viii) Scalars may be brought to the front: Scalars may always be brought to the front of an
expression involving dot products in the sense that if v and w are geometric vectors and
λ is a scalar then
(λv) · w = v · (λw) = λ(v · w) .
We carefully verify this property below.

(ix) Diagonals of a rhombus are mutually perpendicular: A rhombus is a parallelogram in which

all sides have the same length. Below, we give a short proof, using dot products, that the
diagonals of any rhombus are mutually perpendicular.

(x) Altitudes of a triangle intersect in a point: An altitude of a triangle is the line segment from
one of the vertices to the opposite side, such that the line segment is perpendicular to the
opposite side. Below, we apply dot products of geometric vectors to prove that the three
altitudes of any given triangle intersect in a single point.

Examples and explanations:

1. Consider the following vectors:

u = 3i − 4j + 6k , v = 7i − 2k , w = 2i + 5j + k .

(a) Find each of

u · v , u · w , v · w , (u + v) · w , w · w , |w| , (3w − 2v) · (u + 2w) .

(b) Determine whether the angles between each pair of vectors is acute or obtuse. Now
find each of the angles, correct to the nearest degree.

Solutions: (a) We have

u·v = (3i − 4j + 6k) · (7i − 2k) = (3 × 7) + ((−4) × 0) + (6 × (−2))

= 21 + 0 + (−12) = 21 − 12 = 9 ,
u·w = (3i − 4j + 6k) · (2i + 5j + k) = (3 × 2) + ((−4) × 5) + (6 × 1)
= 6 + (−20) + 6 = 12 − 20 = −8 ,
v·w = (7i − 2k) · (2i + 5j + k) = (7 × 2) + (0 × 5) + ((−2) × 1)
= 14 + 0 + (−2) = 14 − 2 = 12 ,
(u + v) · w = u · w + v · w = −8 + 12 = 4 ,
w·w = (2i + 5j + k) · (2i + 5j + k) = (2 × 2) + (5 × 5) + (1 × 1)
= 4 + 25 + 1 = 30 ,
√
|w| = 30 ,
(3w − 2v) · (u + 2w) = 3(w · u) + 6(w · w) − 2(v · u) − 4(v · w)
= 3(−8) + 6(30) − 2(9) − 4(12)
= −24 + 180 − 18 − 48 = 180 − 90 = 90 .

(b) Both u · v and v · w are positive, so the angles between u and v, and between v
and w, are acute (noting in both cases that the angles cannot be zero, because u and
v are not scalar multiples of each other, so are not parallel, and similarly for v and w).

3
 However, u · w is negative, so the angle between u and w is obtuse (and cannot be 180◦ ,
since u and w are clearly not parallel).
Let α be the angle between u and v. Then
u·v 9 9 9
cos α = = √ √ = √ √ = √ ,
|u||v| 32 + 42 + 62 72 + 22 61 53 3233
so ( )
−1 9
α = cos √ ≈ 81◦ .
3233
Let β be the angle between v and w. Then
v·w 12 12
cos β = = √ √ = √ ,
|v||w| 53 30 1590
so ( )
−1 12
β = cos √ ≈ 72◦ .
1590
Let γ be the angle between u and w. Then
u·w −8 8
cos γ = = √ √ = −√ ,
|u||w| 61 30 1830
so ( )
−1 8
γ = cos −√ ≈ 101◦ .
1830
2. Consider the following points in space:

O = O(0, 0, 0) , P = P (1, 2, 3) , Q = Q(−2, 4, −1) .

These points form a triangle OP Q. Find each of the angles at the vertices of this triangle,
to the nearest tenth of a degree (and check that the angles add up to 180◦ ).
Solution: Put
−→ −→ −→
u = OP = i + 2j + 3k , v = P Q = −3i + 2j − 4k , w = QO = 2i − 4j + k ,

so that

u · v = −3 + 4 − 12 = −11 , v · w = −6 − 8 − 4 = −18 , u · w = 2 − 8 + 3 = −3 ,

and
√ √ √ √ √ √
|u| = 1 + 4 + 9 = 14 , |v| = 9 + 4 + 16 = 29 , |w| = 4 + 16 + 1 = 21 .

Let α, β and γ be the angles at the vertices O, P and Q respectively. Then α is the angle
−→ −→
between OP and OQ, that is between u and −w so that
u · (−w) u·w 3 3
cos α = = − = √ √ = √ ,
|u|| − w| |u||w| 14 21 7 6
yielding ( )
−1 3
α = cos √ ≈ 79.9◦ .
7 6

4
 −→ −→
Further, β is the angle between P O and P Q, that is between −u and v so that
(−u) · v u·v 11
cos β = = − = √ √ ,
| − u||v| |u||v| 14 29
yielding ( )
11
−1
β = cos √ √ ≈ 56.9◦ .
14 29
−→ −→
Finally, γ is the angle between QO and QP , that is between w and −v so that
w · (−v) v·w 18
cos γ = = − = √ √ ,
|w|| − v| |v||w| 21 29
yielding ( )
−1 18
γ = cos √ √ ≈ 43.2◦ .
21 29
Note, as a simple check, that

α + β + γ ≈ 79.9◦ + 56.9◦ + 43.2◦ = 180◦ ,

as it should, confirming that we have probably avoided making errors.

3. We explain carefully why the dot product distributes over vector addition.
Proof: Let u, v and w be geometric vectors, so that they have Cartesian forms, say

u = a1 i + a2 j + a3 k , v = b1 i + b2 j + b3 k , w = c1 i + c2 j + c3 k ,

for some real numbers a1 , a2 , a3 , b1 , b2 , b3 , c1 , c2 , c3 . Then

( )
(u + v) · w = (a1 i + a2 j + a3 k) + (b1 i + b2 j + b3 k) · (c1 i + c2 j + c3 k)
( )
= (a1 + b1 )i + (a2 + b2 )j + (a3 + b3 )k · (c1 i + c2 j + c3 k)
= (a1 + b1 )c1 + (a2 + b2 )c2 + (a3 + b3 )c3
= a1 c1 + b1 c1 + a2 c2 + b2 c2 + a3 c3 + b3 c3
= (a1 c1 + a2 c2 + a3 c3 ) + (b1 c1 + b2 c2 + b3 c3 )
= u·w+v·w ,

completing the proof.

4. We explain carefully why scalars may be brought to the front of expressions involving the
dot product.
Proof: Let v and w be geometric vectors, so that they have Cartesian forms, say

v = b1 i + b2 j + b3 k , w = c1 i + c2 j + c3 k

for some real numbers b1 , b2 , b3 , c1 , c2 , c3 . Let λ be any scalar. Then

( )
(λv) · w = λ(b1 i + b2 j + b3 k) · (c1 i + c2 j + c3 k)
( )
= (λb1 )i + (λb2 )j + (λb3 )k · (c1 i + c2 j + c3 k)
= (λb1 )c1 + (λb2 )c2 + (λb3 )c3 = λ(b1 c1 ) + λ(b2 c2 ) + λ(b3 c3 )
= λ(b1 c1 + b2 c2 + b3 c3 ) = λ(v · w) ,

5
 and
( )
v · (λw) = (b1 i + b2 j + b3 k) · λ(c1 i + c2 j + c3 k)
) ( )
= (b1 i + b2 j + b3 k · (λc1 )i + (λc2 )j + (λc3 )k
= b1 (λc1 ) + b2 (λc2 ) + b3 (λc3 ) = λ(b1 c1 ) + λ(b2 c2 ) + λ(b3 c3 )
= λ(b1 c1 + b2 c2 + b3 c3 ) = λ(v · w) ,

completing the proof.

An astute reader might notice that the second half of the proof can be shortened by
using commutativity of the dot product, and invoking the first half of the proof, by
interchanging the roles of v and w :

v · (λw) = (λw) · v = λ(w · v) = λ(v · w) .

5. We explain now why the geometric and algebraic definitions of the dot product of two
vectors produce the same outcome.
Proof: We will start by assuming the dot product is given by the algebraic definition, and
then exploit its algebraic properties. We need to prove

v · w = |v||w| cos θ , (1)

where θ is the angle between v and w. To connect algebra with geometry, we position
the vectors so that they emanate from the same point P , with the tip of v at the point
Q, and the tip of w at the point R, forming a triangle P QR.

Q
b

v v−w

b b
P R
w

Note that the angle θ occurs at the vertex P of the triangle and the opposite side may be
−→
represented by the vector RQ = v − w. By the Cosine Rule for triangles (which follows
from the Theorem of Pythagoras), we have

|v − w|2 = |v|2 + |w|2 − 2|v||w| cos θ . (2)

6
 But we also have, using algebraic properties of the dot product,
|v − w|2 = (v − w) · (v − w)
= v · (v − w) − w · (v − w)
= v·v−v·w−w·v+w·w
= |v|2 − v · w − v · w + |w|2
= |v|2 − 2(v · w) + |w|2 .
Putting this together with (2), we get
|v|2 + |w|2 − 2|v||w| cos θ = |v|2 − 2(v · w) + |w|2 ,
which simplifies, after some cancellations, to the following equation:
|v||w| cos θ = v · w ,
completing the proof of (1). This is enough to establish that the geometric and algebraic
definitions of the dot product coincide.
6. We will now prove that the diagonals of a rhombus are mutually perpendicular, that is,
they intersect at right angles.
b
b

b
b

Proof: Denote the vertices of the rhombus by P , Q, R and S, as in the diagram below:
R
Q b
b

−→
v = PQ

b
b
−→ S
P w = PS
−→ −→
Put v = P Q and w = P S, so that the diagonals are represented by the vectors
−→ −→
PR = v + w and SQ = v − w .
It remains to show that the angle between these vectors is a right-angle, and for this it
suffices to show that the dot product of these vectors is zero. We have

7
 −→ −→
P R · SQ = (v + w) · (v − w)
= v·v−v·w+w·v−w·w
= |v|2 − v · w + v · w − |w|2
= |v|2 − |w|2
= 0,

at the last step because the sides of the rhombus have equal length, that is, because
|v| = |w|. This completes the proof that the diagonals are mutually perpendicular.

7. We can use a similar proof technique to prove that the angle inscribed in a semicircle is
a right-angle.

Proof: Denote the diameter of the semicircle with centre O by P R, and an arbitrary point
−→
on the semicircle by Q. The task is to show that the angle between the vectors QP and
−→
QR is a right-angle.

Q
b

b b b

P v v R
O

−→ −→
Let v be the vector P O, which coincides with the vector OR, since they have the same
length (the radius of the circle) and point in the same direction (along the diameter of
−→
the circle). Let w be the vector QO, which has the same length as v (the radius of the
circle). Then
−→ −→
QP = w − v and QR = w + v ,

8
 so that
−→ −→
QP · QR = (w − v) · (w + v)
= w·w+w·v−v·w−v·v
= |w|2 + w · v − w · v − |v|2
= |w|2 − |v|2
= 0,

since |v| = |w|. This completes the proof that the angle at Q is a right-angle.

8. In this final example, we show that the three altitudes of a triangle intersect in a single
point.
b

b b b

Proof: Label the vertices of the triangle by P , Q and R, the foot of the altitude emanating
from P by B, and the foot of the altitude emanating from Q by C. Denote by S the
point where these two altitudes intersect.

Q
b

A
b

b B
b

b b b

P C R

Draw a line from R passing through S until it meets the side P Q of the triangle. Denote
the point where this line segment meets this side by A. To complete the proof it suffices

9
 to show that the side P Q is perpendicular to the line segment RA, for then RA will be
the third altitude of the triangle, and then all three altitudes will intersect in a single
point. But RS and RA are parallel, so it suffices to prove the following:
−→ −→
RS · P Q = 0 . (3)

We know, however, that P B and QR are mutually perpendicular, as are QC and P R,

which tells us that
−→ −→ −→ −→
P S · QR = 0 and QS · P R = 0 . (4)

We can now calculate as follows, invoking (4) at the fourth and last steps:
−→ −→ (−→ −→) (−→ −→)
RS · P Q = RP + P S · P R + RQ
−→ −→ −→ −→ −→ −→ −→ −→
= RP · P R + RP · RQ + P S · P R + P S · RQ
−→ −→ −→ −→ −→ −→ −→ −→
= RP · P R + RP · RQ − P S · RP − P S · QR
−→ −→ −→ −→ −→ −→
= RP · P R + RP · RQ + SP · RP − 0
−→ −→ −→ −→ −→ −→
= RP · P R + RP · RQ + RP · SP
−→ (−→ −→ −→)
= RP · P R + RQ + SP
−→ (−→ −→)
= RP · P Q + SP
−→ (−→ −→)
= RP · SP + P Q
−→ −→
= RP · SQ
−→ −→
= QS · P R
= 0.

Thus (3) holds, completing the proof that the three altitudes intersect in a single point.

The sequence of steps in the preceding proof is by no means unique, and students can
navigate their own way to the same conclusion. The basic strategy is to expand the dot
product, in order to include points that appear in the equations in (4), and then just use
algebraic properties of the dot product to contract everything down to zero, looking for
any opportunity to use (4) to simplify expressions.

Reference: Easdown, D. (2011). A First Course in Linear Algebra. 3rd Edition. Pearson
Australia, Frenchs Forest, New South Wales.

10