MATHEMATICS INTRO FOR PHYSICS

YORK UNIVERSITY

1. Scalars and Vectors

1.1. Preliminaries. Many physics variables use the real-number system to pro-
vide a quantitative description. Temperature is one such example: on the Celsius
scale we encounter positive and negative real numbers, which allow ordering: we
can always compare two temperature readings and state whether one is higher
than the other (or lower, or equal). A complete description requires specification
of a unit: this distinguishes physics from mathematics. When we adopt the inter-
national unit system SI (meters, kilograms, seconds, Ampere) we can avoid writing
down the unit. However, when doing so we should make clear that we are using
SI to avoid confusion.
In a physics measurement one is limited by the accuracy of equipment. Thus, an
answer like x = 4 m has to be treated carefully. In mathematics the implication
would be that x equals to the integer number 4 (the integers are embedded in
the real numbers). No physics measurement can be as precise as that. Therefore,
we quote results to a certain number of significant digits. A meaningful physics
statement can be: x = 4.00 m - which would imply that we are certain about the
result being above x = 3.995 m and below x = 4.005 m. When calculating answers
to physics problems in this course pay attention to how many significant digits are
provided by the input, use one more digit for intermediate calculations, and quote
your final answer to no more than this one additional digit. Ideally, you will round
the result to the correct number of significant digits.
Some physics variables require two, three, (or perhaps more) real numbers to
specify them. From navigation or geography (GPS) you know that specifying your
location on the surface of the earth requires a longitude and a latitude. In three-
space we locate objects by specifying their (x, y, z) coordinates. Such objects are
called vectors. In this course you will make heavy use of vectors in two dimensions,
and also to a lesser extent in three. In upper-year courses you will encounter higher-
dimensional vector spaces. Some physics variables are represented by matrices, i.e.,
n-by-n arrays of real numbers - again, this you will encounter in second year and
up.
The next section will explain how to deal with objects represented by vectors.
Later an extension of calculus ideas will be provided to show how they apply to
vectorial quantities. This will be needed from the start, since position, velocity,
and acceleration are three-vectors, and they are related by calculus operations.
1
2 YORK UNIVERSITY

The final section will introduce integration as or area-under-a-curve measurement.

This we need, since many physics laws are based on the area under a curve repre-
senting one of the physics variables. The power of physics is to understand these
relationships for relatively simple systems. Modern approaches to many disciplines
(from biology, kinesiology to economics) are driven by understanding laws that are
built on these principles.

1.2. Unit vectors. Vectors can be specified by two different methods. One, more
general, method is to specify them by their length, and by their orientation in
space. When one adopts this method it is clear that they can be moved around
in space and remain the same. The second method depends on a choice of coordi-
nate system. If we defines orientations that specify what we mean by horizontal,
vertical, and depth displacements (such as the edges of a rectangular room we find
ourselves in), then we can represent vectors by a triple of numbers:
~r = (x, y, z) . (1.1)
While this representation depends on a choice of what the (x, y, z) axes are (and
therefore appears to be less general), we nevertheless use it, since it is easy to
work with. In physics one prefers the following notation: one defines the axes of
the coordinate system by unit-length vectors î, ĵ, k̂ and uses vector addition to
assemble the position vector according to:
~r = xî + y ĵ + z k̂ . (1.2)
We can work in the plane by assuming z ≡ 0, and illustrate in Fig (1) what we
mean by the three objects in this case, namely the relevant unit vectors î, ĵ, and
the components of the position vector ~r along the axes (x along î, and y along ĵ).

1.3. Vector addition. The representation of the two-dimensional (2d) position

vector as an addition of contributions along î and ĵ defined a mathematical opera-
tion that has both an algebraic definition and a geometric interpretation. The
algebraic definition of vector addition simply states that we add the x− and
y−components separately for any two vectors ~r1 = x1 î + y1 ĵ, ~r2 = x2 î + y2 ĵ:
~r1 + ~r2 ≡ (x1 + x2 )î + (y1 + y2 )ĵ. (1.3)
The geometric realization of vector addition consists of forming a parallelogram
by moving the arrow representing the second vector such that its tail coincides
with the tip of the first vector arrow. The resulting vector is given by the arrow
connecting the tail of the first with the arrow of the second vector. The order of
which vector to start with doesn’t matter, reversing the order of the geometric
construction completes the parallelogram. Note that this geometric interpretation
consolidates the notion that vectors can be moved in space: they really are defined
by their length (magnitude) and direction.
 MATHEMATICS INTRO FOR PHYSICS 3

Figure 1. A vector in the plane is resolved into x− and

y−components using the î and ĵ unit vectors. The tip of the vector
is at P1 (3, 2) when the tail is put at the origin. If the tail was put
at P2 (2, 0.33), then the same vector ~r would be represented by the
shifted arrow ending at P3 (5, 2.33).

1.4. Vector magnitude. The Cartesian coordinate system defined by î, ĵ, k̂ has
perpendicular axes. The Pythagorean theorem allows us to calculate the length
of a vector given by its components such as in Eq. (1.2). In the plane this is very
obvious from the triangle:
p
r = x2 + y 2 . (1.4)
Note that r is a (positive) scalar. We need to be careful in our notation to distin-
guish between the vector and its magnitude. A physics example: a car is moving
with a speed of 96 km/h. This statement does not say which direction the car
moves in! The statement: a car moves with velocity
~v = 3.0î + 2.0ĵ m/s (1.5)
states specifically the direction of motion. This is a vectorial statement, and
provides a complete understanding of the motion. If this velocity vector does not
change with time, we can understand the motion of the car by tracking the change
in the position and applying this velocity vector for each instant in time.
The speed of the car which moves with the velocity (vector) given by Eq. (1.5)
is obtained by applying the magnitude calculation. We denote the components of
the velocity vector as vx and vy , i.e., ~v = vx î + vy ĵ, and compute in analogy with
Eq. (1.4):
q
v = vx2 + vy2 . (1.6)
4 YORK UNIVERSITY

Figure 2. The magnitude (or length) of a vector is calculated from

its Cartesian components using the Pythagorean theorem.

For our example (1.5) the speed is v = 3.6 m/s. For intermediate calculations you
should keep an additional digit, i.e., it would be prudent to record v = 3.61 m/s.
In fact, the on-line problem-solving environment used in our courses (WebAssign)
assumes that final results are answered with one extra digit (it is not coded properly
for significant digits, and works with a small-percentage window around the ‘exact’
answer to mark your result - sometimes the correctly rounded answer will fall
outside this acceptance window).
1.5. Dot product. Note that you can compare two vectors by their magnitudes.
This is possible, since the magnitudes are real (non-negative) numbers, and these
numbers are ordered. The vectors themselves cannot be ordered. You can have
for two vectors ~r1 and ~r2 the relation x1 > x2 and y1 < y2 , which means that
comparing the two vectors directly is meaningless. You can compare their lengths,
however. A number of physics laws involving vectorial quantities require us to form
products of vectors. The first product, called the scalar product, or dot product
is available in many vector spaces, e.g., in the plane. It associates a scalar number
with any pair of vectors. There are two ways to define it, an algebraic one (in the
Cartesian, or component representation), and a geometric one. Let us start with
the algebraic definition:
~a · ~b ≡ ax bx + ay by . (1.7)
We will avoid dots to denote multiplication of scalars to emphasize that the dot
product is formed from two vectors. The right-hand side of (1.7) is a scalar, which
is a real number without any restrictions (it can be negative). The definition of a
vector magnitude (length) can be traced back to the dot product of a vector with
itself (in which case one obtains the square of the vector length, which is positive).
The dot product (1.7) is easily generalized to more than two dimensions. In this
 MATHEMATICS INTRO FOR PHYSICS 5

course we will use at most N = 3 dimensions, but in special relativity one needs
N = 4, and later (in the context of function spaces, Fourier series, etc.) physicists
use higher N , even N → ∞:
N
X
~a · ~b ≡ ai b i . (1.8)
i=1

Figure 3. The dot product or scalar product of two vectors is a

scalar number (with sign) which measures how much the second
vector points along the first vector, and then scales this number by
the length of the first vector. Since the dot product doesn’t care
about the order of the factors, the interpretation also works the
other way around.

The geometric version of the dot product is quite intuitive:

~a · ~b ≡ ab cos α , (1.9)
where α is the angle between vectors ~a and ~b, and note that on the right hand side of
(1.9) the magnitudes of these vectors appear. In the real plane it is straightforward
to show that (1.8) and (1.9) are equivalent. In vector spaces with N > 3 one can
use the combination of these equations to define the angle between two vectors.
The significance of the dot product in physics, especially apparent in the form
of (1.9) is that it allows to formulate laws between vectorial quantities in which
not only the magnitude of the factors plays a role, but the relative orientation of
the two vectors entering the product. In particular, it is obvious from the relation
cos π2 = 0 that the dot product of two perpendicular vectors vanishes.
6 YORK UNIVERSITY

From the geometric version (1.9) we see that the dot product of two vectors
becomes negative when the angle between them is obtuse (larger than 90 degrees).
To find the angle α we place the vectors tail to tail. A graph of the cosine function
shows that for π2 < α < π the result is negative. A useful geometrical interpretation
of the dot product is that it provides the product of the component of ~b along
~a, which is bcos(α), times the magnitude a. The construction works also the
other way around, since the dot product is symmetric in its two factors. It does
not matter whether the angle α is measured from ~a towards ~b, or the other way
around. This works out, since the cosine function has the symmetry property
cos α = cos (−α). In many other instances we will be careful when measuring
angles, by agreeing that they are positive when increasing in the counter-clockwise
(CCW) direction.
An important physics application of the scalar product can be found when cal-
culating the work ∆W done by a force F~ along some displacement ∆~r. The work
is a scalar quantity and takes into account only that component of the force that
acts along the displacement.

1.6. Vector components and component vectors. We now return to the no-
tion of assembling a vector from its components along the axes to prevent some
confusion. The motion on an inclined plane shown here is a typical problem where
one can simplify matters by a clever choice of coordinate system. In the standard
set-up where î is chosen along the horizontal, and ĵ along the vertical the motion
is two-dimensional, i.e., would require functions x(t) and y(t) for a mathematical
description.
Yet, the motion is one-dimensional along the surface, since the normal force from
the surface cancels the gravitational force component pressing the object into the
surface. Thus, it is clever to choose the direction of motion as one basis vector,
and the perpendicular direction as the other. In order to take advantage of this
trick we need to know how to resolve vectors into components along any possible
direction.
In terms of given î and ĵ any vector ~a can be written as the sum of component
vectors:
~a = ~ax + ~ay = ax î + ay ĵ . (1.10)
Here ~ax and ~ay are component vectors that make up ~a by vector addition, while
ax and ay are scalars, which are called the x− and y−components of ~a. A rather
confusing point is that we might think about ax and ay as being the magnitudes
of ~ax and ~ay . However, this is only true if vector ~a points in the first quadrant
when its tail is at the origin! In general, the components ax and ay can be (and
will be) negative. This leads sometimes to notational confusion: often, when using
trigonometry or geometry to find out relations, we need the magnitudes, rather
than the components, in which case we should, e.g., specify |ax |, or at least provide
a sketch of the situation to make things clear.
 MATHEMATICS INTRO FOR PHYSICS 7

The vector components can be found by using the representation (1.10) and
dotting into the equation (on both sides) with î or ĵ (a trick used often in physics,
so get used to it). This is called taking the projection of the equation along î or
ĵ. For example:
î · ~a = î · (ax î) + î · (ay ĵ) = ax . (1.11)
Here we used the properties that î · î = 1, and î · ĵ = 0. Figure out these properties
using both the algebraic and the geometric versions of the dot product!
1.7. Magnitude and direction from the components. The representation of
vectors in terms of components is easy to work with, but has the drawback that
we need to make a choice of axes (basis vectors î, ĵ, k̂), and specify what they are
when describing ~a in terms of (ax , ay , az ). We stated already in Eq. (1.6) that
the Pythagorean in 2d (and the dot product in 3d) let us calculate the magnitude
a = |~a|. How do we get the direction of the vector?

Figure 4. The direction of a vector is obtained from its compo-

nents. Care has to be taken: the arctan = tan−1 function as im-
plemented on calculators returns answers in the −π/2 < α < π/2
range. Inspection of the signs of the components x and y tells which
of the four quadrants the vector points into. An addition of π or 2
π corrects the answer. The arctan function with a single argument
cannot distinguish between quadrants I and III.

In the plane, and if the vector points into the first quadrant, things are easy.
From geometry/trigonometry we know that
ay
tan θ = . (1.12)
ax
8 YORK UNIVERSITY

Applying the idea of inverse functions, we solve (1.12) for θ. The arctan, or inverse
tan function is implemented on calculators, and formally we write:
ay ay
θ = arctan = tan−1 . (1.13)
ax ax
You should explore this inverse function on your calculator to make sure to un-
derstand that it has nothing to do with 1/ tan x. An inverse functions undoes the
effect of a given function; for another important pair of such functions you can
study ex and ln x.
Things, are, however, a little trickier than just finding the key for arctan on
your calculator. Depending on whether the argument is positive or negative, the
calculator arctan function can only answer in the range 0... π2 , or − π2 ...0, i.e., the
angle will always be in the first or fourth quadrant. Using negative angles to
denote locations in the fourth quadrant is not such a bad thing, in principle (for
significant digits it is much more reasonable to write −0.1◦ , than 359.9◦ ). However,
sometimes vectors do point into the second or third quadrants, and the calculator
is misleading us! Many programming languages provide an arctan function with
two independent arguments to solve this problem. However, in this course you’ll
have to do it yourselves. You have to look at the sign of ax and ay to determine
the quadrant. A negative argument into the arctan function can happen when
EITHER ax OR ay is negative. This happens in quadrants II and IV. The calcu-
lator always goes for quadrant IV, as stated above. You need to add π (or 180◦ )
if ax < 0 and ay > 0. Likewise, the case when BOTH ax AND ay are negative,
your calculator can’t recognize this, since the signs cancel in the division before the
arctan is evaluated. The angle given by the calculator will be in the first quadrant,
and again, you’ll need to correct manually by adding π (or 180◦ ).
It is conventional to measure angles from the positve x−axis as positive in the
CCW sense. Draw a sketch to state your case clearly.

1.8. Vector product or cross product. Midway through the course you will
find this mathematical tool to help you understand three-dimensional space, and
you will learn how to think properly about angular momentum and torque. Later
in magnetism the tool will become indispensable. There is a product of two vectors
which span (define) a plane, that results in a vector, which, however, points in a
perpendicular direction to this plane. The two vectors ~a and ~b will be assumed
to not point in the same (or exactly opposite) directions - otherwise they would
not define a plane. To define this vector (or cross) product ~c ≡ ~a × ~b we need to
specify magnitude and direction. The magnitude is given by
c = ab sin α , (1.14)
where α is chosen such that it falls into the range 0 < α < 180◦ . The latter
guarantees that the above is positive, i.e., describes a magnitude. Concerning the
direction, we have two choices left: perpendicular to the plane defined by ~a and ~b
 MATHEMATICS INTRO FOR PHYSICS 9

could be pointing up or down. To determine it uniquely, we define a right-hand-

rule (left-handers, beware!): align the thumb with vector ~a, the index finger with
~b, and curl the middle finger naturally inside to be at a right angle to thumb and
index finger; the middle finger will give you the orientation of the resulting vector
~c. Doing this trick with the unit vectors î, and ĵ (thumb and index) defines the
z−axis for a right-handed coordinate system, i.e., it implements
k̂ = î × ĵ (1.15)
We never use left-handed coordinate systems in this course: therefore, learn how
to get it right. Convince yourself that using your left hand will result in k̂ would
point in the opposite direction.

Figure 5. The right-hand rule uses the first three fingers spread
out, with the middle finger pointing perpendicularly to the plane
spanned by thumb and index finger to define a right-handed three-
dimensional system of unit vectors. Use of the left hand would
make the z− axis point into the screen - avoid this! The middle
finger is curled naturally inward. This right-hand rule is used to
find the direction for the cross product result ~a × ~b, where ~a and
~b are represented by thumb and index finger, respectively. Order
matters here: ~b × ~a = −~a × ~b !!!

There is a formal cross-product rule for evaluation using determinants. It is

used typically from second-year on.
A geometric interpretation of the magnitude of the cross product is that given
vectors ~a and ~b, and the angle α as measured from ~a towards ~b, then b cos α is the
projection of vector ~b onto the perpendicular direction to ~a (on the same side as
~b). The vector cross product is maximal in magnitude when ~b is perpendicular to
~a.
10 YORK UNIVERSITY

Figure 6. The vector cross product provides a measure of how

much vector ~b contributes in a direction perpendicular to that of
~a; this contribution, b cos α is then multiplied by the magnitude
a = |~a|. Reversing ~a and ~b results in the same magnitude ab| sin α|,
but the resultant vector points in the opposite direction. For angles
π < α < 2π the right-hand rule is awkward to use, but try it out for
α close to 2π to learn how it works! In this case ~b would be below ~a
in the figure.

2. Calculus
2.1. Preliminaries. Our textbook works at the pre-calculus level. Yet, the course
is supposed to be calculus-based. Why do we do this, and how can we put the
calculus in? There are multiple reasons for our choice. The main reason is that
the physics laws are valid and can be formulated without taking limits. The
formalism of calculus was developed by Newton, Leibnitz and others to derive nice
symbolic results in mechanics. The physics principles, such as Newton’s laws, the
gravitational law, or the laws of electricity and magnetism by themselves have to
be valid when using finite intervals. Finite-interval arithmetic is used commonly
in computer programs to solve problems for which a symbolic answer cannot be
found using Calculus tools. Thus, the pre-calculus version of physics has acquired
some modern validation.
The main problem with so-called calculus based texts is that very quickly they
move from simple ideas such as ‘the slope of a curve is given by the first deriva-
tive’ to ‘calculate the area under the curve’. This happens very early in classical
mechanics, yet in mathematics it is covered later under integral calculus (and not
 MATHEMATICS INTRO FOR PHYSICS 11

covered in your typical high-school calculus course). To make things worse, one
needs quickly a new type of integral (line or curve integral to calculate the work
done by a force acting on a moving object in two or three dimensions), then the
moment of inertia (a volume integral with potentially complicated boundaries).
In electricity and magnetism we need another type of integral to measure how a
force field permeates some surface. It doesn’t stop there: we also want to solve
differential equations, and we even write down some differential equations with
partial derivatives (the wave equation) before the course is over. Past experiences
with a deeply conceptual book that is based on calculus (by Randy Knight) have
generated student complaints ‘why are we doing new mathematics all the time?’.
Modern books addressing College Physics (college in the USA is not a commu-
nity college, but really another name for university, usually without a graduate pro-
gram) have found a way to emphasize the important laws without over-burdening
the student with advanced mathematics (which is really covered only in the second
year of university mathematics). The book by N. Giordano matches Knight’s text
in emphasis on the concepts. Very elegantly, it deals with topics such as Gauss’ law
in electricity without the mathematical burden. The present summary of calculus
ideas is provided to put the calculus into the course where it is most appropriate
and straightforward, namely the mechanics part. The first-year student in On-
tario (and most likely elsewhere in Canada) is ready for this, since it really relies
only on high-school calculus as a prerequisite. In a very straightforward way does
this chapter introduce the idea of how area is measured under a curve. It also
reviews some of the trigonometric and the exponential function to highlight their
derivative properties - something that has been eliminated from the high-school
curriculum when grade 13 (or OAC) was finally eliminated.
Let us make just one important comment here: the reason why physicists work
liberally with objects such as ∆t, and even treat the symbol dt as if it were a finite
∆t, is that physics measurements (or physical reality) only exists for finite time
intervals ∆t = t2 − t1 . Below a certain scale, the notion of classical mechanics
begins to fail, the boundaries of a material object become fuzzy, since the laws
of quantum physics take over. It is a beautiful result that in reality one can
reduce ∆t by orders of magnitude when defining the instantaneous velocity of an
object while finding that v( t1 +t
2
2
) = xt22 −x
−t1
1
becomes independent of t2 − t1 , but a
measurement cannot take the limit ∆t → 0. Thus, physical reality cannot depend
on the question whether a limit exists or not. The mathematical convenience of
expressing results, however, does depend very much on this ability.

2.2. Derivative. In one-dimension the position of a particle is completely de-

scribed by the function x(t). In kinematics we study how this function is related
to the variables that control its behaviour, namely the velocity v(t), and the ac-
celeration a(t). In dynamics we are trying to answer the question how the net
force acting on the particle at any location x governs the behaviour of x(t) based
12 YORK UNIVERSITY

upon two initial conditions that make the motion unique. For a given time in-

Figure 7. The trace of position versus time can be used to compute

the average velocity for a given time interval. In the limit of van-
ishingly small time interval, ∆t → 0, this leads to the instantaneous
velocity defined as the first derivative of x(t) which is represented
by the slope of the graph at time t. In physics measurements ∆t re-
mains always finite, but for sufficiently small ∆t the velocity doesn’t
change anymore within the significant number of digits.

terval ∆t = t2 − t1 , which we control by using, e.g., a stopwatch we measure the

displacement ∆x, which is given by ∆x = x(t2 ) − x(t1 ) ≡ x2 − x1 . The average
velocity for this time interval then is defined as
∆x
vavg ≡ . (2.1)
∆t
Consider now the example of a car starting from a stop with the throttle fully
depressed and moving along a perfectly straight road. We expect the speed of the
car to increase rapidly at first, and to level off eventually, as the maximum speed
is approached. Since the motion is in one dimension, and never changes direction
speed and velocity will be the same (we choose the direction of motion as the
positive x-axis). The velocity itself is a function of time. The graph displaying
v(t) becomes boring after t0 when the top speed is reached. The time t0 will be on
the order of a minute. If we look at the part of the curve where v(t) changes with
time we expect that the average velocity will be different if we choose ∆t = 2 s vs
∆t = 10−2 s = 10 ms. Why? Our intuition tells us that on the 2 s time scale the
car is accelerating, and enough time has passed to observe a measurable change
in velocity before and after this time interval. When we reduce the time interval
 MATHEMATICS INTRO FOR PHYSICS 13

significantly below 1 s, however, the change in velocity ∆v will become rather

small, and we are measuring the instantaneous velocity at time t:
∆x dx
v(t) ≡ lim ≡ . (2.2)
∆t→0 ∆t dt
Geometrically, we observe that the velocity can be interpreted as the rate of
change of position with time, or as the slope in the x(t) graph shown below. The
slope in the v(t) graph given above represents the instantaneous acceleration a(t).
Initially, it is approximately constant: the engine delivers a maximum forward
thrust (actually a backward thrust by the drive wheels against the ground, and a
reaction force from the ground propels the car forward). This thrust is opposed
by small frictional forces in the bearings. Eventually, air resistance builds, since
the car picks up speed. When the velocity reaches its maximum value all of the
forward thrust is compensated by air drag (and the small frictional force from the
bearings).
Differential calculus is sufficient to understand how to obtain velocity from po-
sition, and acceleration from velocity. Our ultimate goal, however, is to do the
reverse: to start with acceleration at any time t, and move towards determining the
velocity change ∆v over some time interval ∆t, and from the determined v(t) then
reconstruct the displacement ∆x for some ∆t. Why is this our goal? Newton’s
second law relates the net force acting on a particle of mass m to the acceleration.
Thus, if we understand which forces act on a particle at any position x, then we
know the acceleration, and from it we can then solve for the motion. This is the
ultimate goal of dynamics, and completes the understanding of how motion of a
particle comes about. Before we move on to explain the opposite operation to
differentiation, we make some remarks about how the differential calculus ideas
are generalized in the case of multi-dimensional motion.
2.3. Derivatives of vectors. Given a vector that depends on a parameter (time
in our example), ~r(t), we can generalize Eq. (2.2)
∆~r
~v (t) ≡ lim , (2.3)
∆t→0 ∆t

using the displacement vector ∆~r. We can reduce these expressions to known
quantities by using the Cartesian representation:
dx dy dz
~v (t) = î + ĵ + k̂ , (2.4)
dt dt dt

We have the equivalent of the one-dimensional mathematics in all three compo-

nents, since
~v (t) = vx î + vy ĵ + vz k̂ . (2.5)
Therefore, we need to apply our high-school calculus ideas to the one-dimensional
motions x(t), y(t), (and possibly z(t)), and then combine it to get the full picture.
14 YORK UNIVERSITY

Figure 8. The derivative of a vector with respect to a parameter is

defined in terms of the derivatives of the component functions x(t)
and y(t). The graph illustrates this for finite time interval ∆t, before
the limit is taken. The unit vectors î and ĵ are fixed in time for a
non-moving coordinate system.

Whether it will be easy to carry out the program or not will depend on the forces
acting on the particle, since they may depend on all three coordinates x, y, z in
some non-trivial way.
2.4. Area under a curve. We showed that the relationship between velocity
and position follows from the definition of velocity; when looking at the velocity
graph vs time, it should tell us what the slope is at any time in the corresponding
position vs time graph. How can we figure out an x(t) graph from a given v(t)
graph?
Let us begin from Eq. (2.1) and solve it for the displacement (while dropping
the avg label by assuming that ∆t is sufficiently small):
∆x = v(t)∆t . (2.6)
Let us assume that the particle is located at time t at position x(t), at which it
has velocity v(t) which will not change appreciably over the short interval ∆t. At
the end of this interval the particle is located at x(t + ∆t) = x(t) + ∆x. From
the graph shown for a linearly increasing v(t) (which corresponds to constant-
acceleration motion for some larger time interval) we can infer that the statement
x(t + ∆t) = x(t) + v(t)∆t (2.7)
 MATHEMATICS INTRO FOR PHYSICS 15

Figure 9. The velocity as a function of time for constant-

acceleration motion. The particle starts at t = 0 with velocity v0 .
The displacement of the particle is observed to be the area between
the trace of v(t), the t−axis and the vertical lines t = ti = 0, and
t = tf . The area is made up of two parts: a rectangle corresponding
to the contribution from the initial velocity, and a triangle which is
due to the constant acceleration.

establishes that the displacement is given by the area under the curve v(t) for the
segment t → t+∆t. More precisely, the area bounded by the vertical lines t, t+∆t,
by the t-axis and by the trace of v(t) yields the displacement. Eq. (2.7) makes it
clear that we start from a given x(t), and the area law tells us to which position
we advance after the time ∆t. Wait, you might think, there is an inaccuracy here:
the area under the curve is a rectangle plus a triangle, whereas you only accounted
for the rectangle part! The small rectangle is of higher order, it is proportional to
∆t∆v, and will drop out in a limiting process.
The same area-under-a-curve trick will work to get the velocity function v(t)
from a known a(t). Solving for the motion from a given acceleration function
a(t) involves the repeated application of ‘integration’. Let us illustrate this by the
constant-acceleration example known as a ‘physics law’ from high school, which
merely becomes an almost trivial example in university:
We start with known acceleration a = const. The area under the horizontal line
from fixed initial time t0 to arbitrary time t is found to be

v(t) = v(t0 ) + a(t − t0 ) . (2.8)

16 YORK UNIVERSITY

Figure 10. A similar example to the previous figure, except that

the initial velocity is negative. For the choice of v(t) and time in-
terval tf − ti the contributions from the initial velocity and from the
acceleration just cancel. The area law requires to measure the area
with sign which is exactly what definite integration does. The par-
ticle was moving to the left at t = 0, comes to a stop at t = 5 s due
to the positive acceleration, and returns to the initial position at
t = 10s.

Note that the area-law does not tell us what v(t0 ) is! It only tells us how the
velocity changes from the value at time t0 . Since we never use absolute time,
we could set our stopwatch to t0 = 0. Then Eq. (2.8) for a given a value would
represent a family of curves that start at t = 0 at any height v(0) = v0 , and are
straight lines with slope a. We would state the answer as v(t) = v0 + at. The
meaning of v0 is that of an initial condition. We can start the free-fall motion
of a rock by releasing it from our hand (v0 = 0), or by tossing it (up or down),
therefore giving it some initial velocity.
Let us pick one such choice in the graph (with a > 0 and v0 > 0). The area
under this curve should give us the displacement of the particle. The area between
t0 = 0 and arbitrary time t, the t-axis and the graph consists of two parts: a
rectangular piece of area v0 t, and a triangular piece given by ‘one half times base
times height’, i.e., 21 t(at). We arrive at

1
∆x = v0 t + at2 . (2.9)
2
 MATHEMATICS INTRO FOR PHYSICS 17

Again, the area law has nothing to say about an initial condition, namely the
initial position x0 = x(t0 ) = x(0), and the full answer becomes
1
x(t) = x0 + v0 t + at2 . (2.10)
2
You can verify that differentiation of (2.10) results in the velocity (2.8), and further
differentiation in the statement that the acceleration is a constant function of time
a(t) = a.
Area-under-a-curve laws appear in other branches of physics in various contexts.
They also appear in mathematical descriptions of biology, economics, engineering,
or other fields where physicists have pioneered a deeper understanding of concep-
tual relationships. The calculus challenge of re-tracing the steps that lead to x(t)
when we have some function a(t) requires the notion of integrals or anti-derivatives.
It is needed when areas can no longer be obtained from simple geometric shapes.
Note that the area formula for a circle is derived in calculus. In high school you
were simply using a given formula A = πR2 (a rabbit out of a hat?).
2.5. Anti-derivative. Given our simple example, where the position was given
by a general quadratic function (2.10), and the velocity by a linear function (2.8)
where we noted that the velocity follows by differentiation of the position function,
we should be able to reverse the process. Thus, in general, when we try to find a
position function that corresponds to a given velocity function, we are looking for
the function whose derivative is v(t). In integral calculus you will find that this is a
hard process, and that simple functions can have complicated (or unknown) func-
tions as their anti-derivatives. Sometimes they are sufficiently important that we
call them ‘special’, study their properties and equip calculators and programming
languages to provide easy access to them.
In integral calculus you prove the important theorem that if f (t) and F (t) satisfy
f (t) = dFdt
, then the area under the curve f (t) from t1 to t2 (between the vertical
lines) is given by F (t2 )−F (t1 ). Note that area is computed with sign here, positive
and negative areas (by being above and below the t-axis) do cancel. In this course
you may have to apply this rule occasionally. In upper-year courses you will use
it on many occasions.
In mathematics you prove what the derivative functions are for some elemen-
tary functions. We need these properties in first-year physics for sine, cosine, the
exponential function (apart from the simple power law). While you should com-
mit some of this to memory, you may be given a table. sin0 = cos, cos0 = − sin,
exp0 = exp, where the prime denotes differentiation with respect to the argument.
If the velocity is given by the sine function, then the position will look like an
inverted sine function (plus a constant).
2.6. Chain rule. The differentiation rule for composite functions is of utmost
importance in physics, yet it was perhaps practiced too little in your mathematics
eduction. A composite function can be denoted as f (g(t)), i.e., your ‘outermost’
18 YORK UNIVERSITY

Figure 11. The trace of the velocity over time for a particle re-
leased by a compressed spring: the particle starts from rest, in-
creases its velocity, reaches a maximum velocity when the spring
is in its natural state (equilibrium), then it starts expanding the
spring and slows down. The area under the curve which yields the
displacement is approximated by a sum obtained from sampling the
particle velocity at ∆t = 0.5 s intervals. The sum of the rectangles
will approach the area under the curve when the sampling rate is
increased, i.e., the interval ∆t will be decreased towards zero. In
integral calculus this area is computed from the anti-derivative.

function has as its argument√itself a function of the independent variable. An

example could be f (t) = cos ( t2 + 1). How do we find the derivative with respect
to t? The rule is
df df dg
= . (2.11)
dt dg dt
The proof is in the calculus course. We just need to know how to use it. For
our example it means: use the cosine derivative pretending the argument was the
independent variable. Then multiply by the derivative of the inner function with
respect to t.
This rule will be needed even in simple examples in physics for the following
reason: one cannot calculate functions of dimensionful variables, i.e., sin t, or
exp t are OK in mathematics, but not in physics. Simple power-law functions
are an exception to this rule. The simplest combination of two powers, such as
t0 + t1 = 1 + t shows that in physics this is a forbidden operation, unless one makes
the convention that t is a number in some unit system. In general, the statement
 MATHEMATICS INTRO FOR PHYSICS 19

Figure 12. The trace of an oscillator solution (a mass attached to a

spring is swinging to the right and left without damping) in a physics
representation in SI units, as well as in dimensionless units. The
length is scaled out using the amplitude x0 , while time is measured in
units of the oscillation period T . The argument of the sine function
has to be dimensionless (an angle measured in radians). The velocity
calculation requires the chain rule, and it is this factor that ensures
that the derivative has dimensions of length divided by time.

f (t) = 1 + t would mean that we are adding a dimensionless number (here 1) to

the variable t, i.e., time which is dimensionful. take f (t) = t + t2 : what would be
the meaning of adding seconds to seconds squared???
For this reason functions such as sine, cosine, exp must have arguments in physics
that are made dimensionless. We may have some time scale t0 , and a meaningful
function exp (−t/t0 ). In circular and oscillatory motion we will use frequency ν or
circular frequency ω = 2πν, or the period of the motion T = 1/ν = 2π/ω to apply
meaningfully:
cos ωt = cos (2πνt) = cos (2πt/T ) . (2.12)
While differentiating such expressions (or finding anti-derivatives) the chain rule
will bring out dimensionful factors, which are needed, in fact, to turn a function
describing a position into a velocity, etc.

2.7. Functions defined by their derivative properties. Laws in physics (and

in other disciplines) are often relating variable functions to their derivatives (rate
of change), or anti-derivatives. Answers to such laws (solutions) are then functions
of the independent variable. We will look at three functions that come up in first-
year physics (and also in chemistry and even biology). The class of exponential
20 YORK UNIVERSITY

functions is defined by the relationship

df
= f (t) . (2.13)
dt

Figure 13. The exponential function is used in science to model

growth that depends on the population size. The rate of change
(slope) for any value of the argument is equal to the function value.
Instead of using Euler’s number e ≈ 2.7 as base one could define
the function at/τ , e.g., with a = 2. For this function shifting the
argument by one unit to the right would correspond to doubling the
population instead of multiplying by e ≈ 2.7.

Here t is a mathematical parameter (it could be time in dimensionless units),

and a simpler way to write the equation is f 0 = f . Why are we interested in a
function with such a property? The equation states that for any value of t the rate
of change of the function (the slope of the graph) equals the function value. One
of the most fundamental relations in population growth is the statement that the
replication rate is proportional to the population. This assumes that there is space
and food for unlimited growth. A bigger population results in more offspring. The
population is measured by the number N (t), and we state that how this number
 MATHEMATICS INTRO FOR PHYSICS 21

changes with time depends is given by

dN
= aN (t) . (2.14)
dt
If t was dimensionful, we might argue that the dimensions of the proportionality
constant a should be t−1 to make things right. Let us, however, write the equation
in a physically meaningful way: we introduce a time constant τ , characteristic of
the process we are investigating. The law is then written as
df t
= af ( ) . (2.15)
dt τ
If we write this in terms of dimensionless time ξ = t/τ , i.e., time is measured
as a multiple of the time constant τ , we see that the right hand side of (2.15) is
df dξ
af (ξ), while the left-hand side by the chain rule becomes dξ dt
. The second factor
dξ 1
is simply dt = τ . We thus arrive at
df
= τ af (ξ) . (2.16)
dξ
Therefore, the proportionality constant in the population growth model (2.14) a =
1/τ is expressible as the inverse of a time constant. The mathematical statement
(2.13) is for the special case when τ = 1 unit. What is the meaning of this time
constant τ ?
The mathematical statement (2.13) does not determine a function uniquely, it
defines the ‘class’ of exponential functions. To select one of these functions from
the class one has to supplement (2.13) with a statement of what the value of f is
at some argument. All exponential functions pass through the point f (0) = 1. We
need an additional selector. In physics (and most other science applications) we
use a strange number to select our preferred f , namely Euler’s number e ≈ 2.7183,
which is an irrational number defined by a limit (a transcendental number such as
π, since it does not follow from an algebraic equation). When looking at traces of
electric signals on an oscilloscope we will be happy with the approximation e ≈≈ 3.
Exponential growth is characterized by the (dimensionless-time) function exp(t) ≡
t
e , which means that as we move along the time axis by one unit the population
increases by a factor of almost three. This is true whether we go from t = 0 to
t = 1 (unit), from t = 1 to t = 2, or anywhere on the graph. Thus, the meaning
of the time constant τ = 1/a is clear now: it is the length of time over which the
population grows by a factor of 2.718..., or roughly a factor of 3.
We can also discuss radioactive decay: the number of decays observed per unit
time is proportional to the number of parent nuclei still present in the source:
dN
= −aN (t) . (2.17)
dt
The negative sign is needed, since the rate of change is negative: after some time
τ the number of undecayed nuclei will be less; for a finite time interval ∆t this
22 YORK UNIVERSITY

would be expressed as ∆N = Nfin − Nin , and since Nfin < Nin we see that the
derivative on the left-hand side of (2.17) is negative. In dimensionless units the
function describing this behaviour is
exp(−t) ≡ e−t , (2.18)
and the meaning of the time constant τ = 1 unit in this case is that it takes time
τ for the population to drop by almost a factor of three.

Figure 14. Exponential decay is a process where the (negative)

rate of change of some quantity is proportional to the amount of
quantity left. Within one characteristic lifetime the amount is re-
duced by a factor of e−1 ≈ 0.37. For each further passage of a
lifetime the amount is again reduced by the same factor.

We will use the exponential function to describe how mechanical energy is con-
verted in a real-life oscillator into heat (the energy loss is proportional to the
motional energy), or how electrical oscillations die down due to damping (elec-
trical energy is converted into heat in a resistor), The oscillations themselves are
described by a class of functions that satisfy the law (again in dimensionless time):
d2 f
= −f (t) . (2.19)
dt2
Both the sine and cosine functions solve this equation, based upon their relation-
ships under differentiation: sin0 = cos and cos0 = − sin. The details of this relation
 MATHEMATICS INTRO FOR PHYSICS 23

are discussed in the lectures on oscillations, but the basic properties are needed to
understand the mathematical description of circular motion.

2.8. Complex numbers. Hundreds of years ago mathematicians wondered about

the question why some polynomials of order n have n roots in the real numbers,
while others didn’t. It happens already for the quadratic. The equation
z2 + 1 = 0 (2.20)
has no solution for real z. The idea to extend the real-number system to complex
numbers by defining the imaginary unit i2 = −1 (leading to two solutions to
(2.20), namely z = ±i) leads to a number field that is closely related to the two-
dimensional real plane. The complex number
z ≡ x + iy (2.20)
is defined in terms of independent real x and y. The beauty of this number system
is that any nrmth order polynomial has exactly n roots. While physics variables
to be measured have to be representable by real numbers (there is no meter in
this world that can show a complex number for an answer!) complex numbers
are used in physics. It is a convenience in electricity (electrical engineers really
use them!) and a necessity in quantum mechanics - even though the measurable
quantities come out as real-valued. We simply end this sub-section by stating a

Figure 15. The graphical representation of the complex number

plane illustrates that the real and imaginary parts of a complex
number are independent of each other, very much like the Cartesian
x− and y−components of a vector in the plane.

powerful relationship, namely Euler’s theorem which shows the relation between
24 YORK UNIVERSITY

the exponential and basic trigonometric functions (which you might have wondered
about in the previous subsection):
eiα = cos α + i sin α (2.21)
2.9. Taylor series. It is possible to verify Euler’s theorem (2.21) by observing
the Taylor series expansions for the three functions involved:
t3 t5 t7
sin t ≡ t − + − ± ... , (2.22)
3! 5! 7!
where the factorial n! ≡ 1 · 2 · 3 · ... · n, and
t2 t4
cos t ≡ 1 − + ± ... , (2.23)
2! 4!
while
t2 t3 t4
exp t ≡ 1 + t + + + + ... (2.24)
2! 3! 4!
The proof of these expansions comes late in your Calculus education, since it is an
application of integral calculus. The meaning, however, is important in physics:
a function can be approximated locally by starting with the function value, and
having some derivative information for the local point t = t0 :
f 0 (t0 )h f 00 (t0 )h2 f 000 (t0 )h3
f (t0 + h) = f (t0 ) + + + . (2.25)
1! 2! 3!
The main idea of the Taylor series (2.25) is to replace a complicated function f by a
polynomial. Eqs. (2.22-24) state that relation for the three ‘complicated’ functions
used in this course using t0 = 0 as the expansion point. Computer arithmetic only
knows elementary operations (add, multiply and their inverses). Calculators and
floating-point units in CPUs use expansions based upon (2.25) to compute numbers
for a given argument t. When modelling nature we use expansions such as (2.22-
25) to simplify laws. An example in this course is when the pendulum equation is
reduced to a simple oscillator for small-amplitude oscillations.
This summary of what will be used in the first-year courses should serve as
motivation to study mathematics ideas in some depth. It is meant as a shortcut
so that one can apply them without being help up by the (sometimes tedious)
process of proving them.

mh, Aug 2011