Space and Time in Special Relativity

Dan Styer;
c 28 September 2015; 8 September 2017

Introduction. You’ve studied space and time in special relativity before. This essay simply gives you a
refresher and a new perspective.

The basic idea is this: Common sense — going back to Galileo — tells us that any inertial reference
frame is as good as any other, and experiment (for example the classic experiments by Trouton and Noble)
backs up that common sense.

But experiment also shows that the speed of light is the same in all inertial reference frames, which is
certainly not in accord with common sense.

Now speed is defined as

distance traveled
speed = .
time elapsed during that travel
So if speed doesn’t behave in the common-sense way then it must be because (a) time doesn’t behave in the
common-sense way, or (b) distance doesn’t behave in the common-sense way, or (c) both.

Lorentz transformation

Situation. Inertial frame F’ moves at constant speed V relative to inertial frame F, and the two frames
coincide at time t = t0 = 0.

event
y y'

F F'
x x'

z z'

An event has space-time coordinates (x, y, z, t) in frame F and space-time coordinates (x0 , y 0 , z 0 , t0 ) in frame
F’. How are the two sets of coordinates related?

Transverse coordinates. It’s clear that y 0 = y and z 0 = z. It’s equally clear that the coordinates y, z will
have no effect on the coordinates x, t.

Linearity. If a ball moves with constant speed in one frame (x = A + Bt) then it moves with constant
speed in the other frame (x0 = C + Dt0 ). From this, linear algebra books prove (it’s surprisingly difficult)

1
that for any event

x0 = b11 x + b12 t
0
t = b21 x + b22 t.

The coefficients b11 , b12 , b21 , and b22 depend on the frame speed V , but do not depend on x or t: they are
the same for all events.
I like to write this relation as

x0 = a11 (β)x + a12 (β)ct (1)

ct0 = a21 (β)x + a22 (β)ct. (2)

where β = V /c is dimensionless and ct has the dimensions of x. (Some people like to write the product ct
as x0 , but that strikes me as overkill at this point.)
Our job now is to find these four functions a11 (β), a12 (β), a21 (β), and a22 (β).
Use uniform motion of frames. If x0 = 0, our situation says that x = V t or, what is the same thing,
x = βct. But equation (1) says that if x0 = 0, then
a12 (β)
x=− ct.
a11 (β)
We conclude that
a12 (β) = −βa11 (β).

If x = 0, our situation says that x0 = −V t0 or, what is the same thing, x0 = −βct0 . But dividing
equation (1) by equation (2) says that if x = 0, then
x0 a12 (β)
= .
ct0 a22 (β)
We conclude that
a12 (β) = −βa22 (β),
which, combined with last paragraph’s result, gives

a11 (β) = a22 (β).

In summary, the transformation equations must have the form

x0 = a11 (β)(x − βct) (3)

0
ct = a21 (β)x + a11 (β)ct. (4)

Use invariance of light speed. The speed of light is the same in both frames. That is, if x = ct then
x = ct0 . For events on the light front, equations (1) and (2) become
0

ct0 = a11 (β)ct + a12 (β)ct

0
ct = a21 (β)ct + a22 (β)ct.

2
The immediate consequence is that a11 (β) + a12 (β) = a21 (β) + a22 (β) which, combined with previous results,
gives a12 (β) = a21 (β).
In summary, the transformation equations must have the form

x0 = a11 (β)(x − βct) (5)

0
ct = a11 (β)(−βx + ct). (6)

Use inverse transformation. We’ve been talking about how to transform coordinates in frame F to
coordinates in frame F’. But of course frame F’ is just as good a frame as frame F. So to convert from
coordinates in frame F’ to coordinates in frame F, just use the same equation except replace β with −β:

x = a11 (−β)(x0 + βct0 ) (7)

0 0
ct = a11 (−β)(βx + ct ). (8)

Well of course we could also solve the two equations (5) and (6) algebraically for x and ct. The algebra is
straightforward and results in
x0 + βct0
x =
a11 (β)(1 − β 2 )
βx0 + ct0
ct = .
a11 (β)(1 − β 2 )
Comparing these two equation pairs demonstrates that
1
a11 (β)a11 (−β) = . (9)
1 − β2

Use isotropy of space. Here are two questions:

(1) A clock moves at speed V to the east past a standing person. When the clock passes the person, it
reads 0 seconds. What time has passed, in the person’s frame, when the clock strikes 14 seconds?
(2) A clock moves at speed V to the west past a standing person. When the clock passes the person, it
reads 0 seconds. What time has passed, in the person’s frame, when the clock strikes 14 seconds?
These two questions differ only by the substitution of “west” for “east”, and since we don’t expect the
direction of the clock’s motion to make any difference, we expect these two questions to have the same
answer. You might call this “east/west symmetry”, or “right/left symmetry”, or “reflection symmetry”, or
even by the high-falutin’ name of “isotropy of space”, but what you call it matters not at all. What matters
is that the two different experiments are described differently:
(1) The person stands at the origin of frame F, the clock rests at the origin of frame F’. The event “clock
strikes 14 seconds” has coordinates x0 = 0, t0 = 14 s, so (by equation 8), the answer to the question is
t = a11 (−β)(14 s).
(2) The person stands at the origin of frame F’, the clock rests at the origin of frame F. The event
“clock strikes 14 seconds” has coordinates x = 0, t = 14 s, so (by equation 6), the answer to the question is
t0 = a11 (β)(14 s).

3
 East/west symmetry demands that these two questions have the same answer, so a11 (β) = a11 (−β) and
equation (9) results in
1
a11 (β) = p . (10)
1 − β2

To conclude, this argument has determined that

x − βct
x0 = p (11)
1 − β2
−βx + ct
ct0 = p . (12)
1 − β2

This pair is called the “Lorentz transformation”. (Why this centerpiece of Einstein’s relativity theory is
named after Dutch physicist Hendrik Lorentz is a question of history not to be pursued here.)

We will usually write these as

x−Vt
x0 = p
1 − (V /c)2
t − V x/c2
t0 = p .
1 − (V /c)2
If there are two events, then they are separated by
∆x − V ∆t
∆x0 = p
1 − (V /c)2
∆t − V ∆x/c2
∆t0 = p .
1 − (V /c)2

Problem 1: We traced through the logic by first demanding the invariance of light speed, then demanding
that inverse transformation works correctly. Work through the derivation again but this time invert those
two demands.

Acknowledgments: Rob Owen made suggestions concerning the “isotropy of space” step, and my students
in the fall semester 2016 offering of Physics 212, Modern Physics, refined the arguments of that same step.

4
Speed transformation

Well, if space and time transform between frames through the non-common-sensical Lorentz transformation,
and if speed is just the quotient of a space with a time, then how do speeds transform?
A bird flies with uniform speed along the x-axis. Two events happen along the bird’s trajectory: perhaps
(1) it raises its wings and (2) it lowers its wings. In frame F these two events are separated by distance ∆x
and time ∆t; in frame F0 they are separated by distance ∆x0 and time ∆t0 . The speed of the bird is vb in
frame F and vb0 in frame F0 , where
∆x
vb =
∆t
and
∆x0
vb0 = (use Lorentz transformation to find. . . )
∆t0
∆x − V ∆t
= (divide numerator and denominator by ∆t to find. . . )
∆t − V ∆x/c2
vb − V
=
1 − vb V /c2

Examples:
common sense correct relativistic formula
vb = 100 mph V = 20 mph vb0 = 80 mph vb0 = 80.000 000 000 000 02 mph
vb = c V 6= c vb0 =c−V vb0 = c
vb = − 34 c V = 34 c vb0 = − 32 c vb0 = − 24
25 c
Problem 2: Apple on Train. A train moves east at speed vt relative to the earth. Someone within
the train tosses an apple east at speed va relative to the train. In the earth’s frame the apple must move at
some speed faster than the train. Call this “excess speed” by the name vax .

a. Show that
1 − vt2 /c2
vax = va .
1 + va vt /c2
b. This formula shows that vax is always less than va . Justify this fact qualitatively using the concepts of
time dilation and length contraction.

c. If vt → c, what does vax approach?

Time dilation, length contraction, and the relativity of synchronization

Start with Lorentz transformation

x−Vt
x0 = p (13)
1 − (V /c)2
t − V x/c2
t0 = p . (14)
1 − (V /c)2

5
If there are two events, then they are separated by
∆x − V ∆t
∆x0 = p (15)
1 − (V /c)2
∆t − V ∆x/c2
∆t0 = p . (16)
1 − (V /c)2
Further, frame F0 is just as good an inertial frame as frame F, so we can transform from coordinates in F0
to coordinates in F simply by replacing every V with a −V :
∆x0 + V ∆t0
∆x = p (17)
1 − (V /c)2
∆t0 + V ∆x0 /c2
∆t = p . (18)
1 − (V /c)2

To derive time dilation, think about what time dilation means: The single moving clock ticks twice —
two events. The clock is stationary in frame F0 , so these two ticks are separated by ∆x0 = 0 and ∆t0 = T0 .
In frame F, the time elapsed T is given by equation (18), so
T0
T =p .
1 − (V /c)2

To derive length contraction, think about measuring the length of a rod moving in the laboratory:
arrange for two events, simultaneous in the laboratory frame (∆t = 0), to occur at the two ends of the
moving rod. These events are separated by length L = ∆x in the laboratory frame. In the rod’s frame F0 ,
the two events are not simultaneous, but they don’t need to be: the distance between them is ∆x0 = L0 ,
regardless of ∆t0 , because the rod is at rest. Equation (15) gives the relationship
L
L0 = p
1 − (V /c)2
which is usually written
p
L = L0 1 − (V /c)2 .

The third principle is relativity of synchronization. A pair of moving clocks ticks simultaneously
(∆t0 = 0) in their own frame (F0 ), and the distance between them in that frame is L0 = ∆x0 . In the
laboratory frame the ticks are not simultaneous: according to equation (18), those two ticks are separated
by a time
V L0 /c2
p .
1 − (V /c)2
But just because those two ticks are separated by this much time doesn’t mean that this is the difference in
the times announced by these two clocks: The two clocks are ticking slowly (time dilation) so the difference
p
in time announced is the smaller time 1 − (V /c)2 times the above, namely
L0 V
.
c2
For a pair of moving clocks (synchronized in their own frame) the rear clock is set ahead.

6
Pole in the barn

Most barns have two doors, so that you can pull a trailer into the barn, stop and unload it, and then drive
out again without backing up. The farm where I grew up in Pennsylvania had a barn extending exactly
100 feet between its two doors.

One day a world champion pole vaulter came to visit our farm. He carried his favorite pole which, by
coincidence, was also exactly 100 feet long. The champion boasted that he was so fast that, even carrying
his pole horizontally, he could run right through our barn at the speed of V = 45 c.

“At that speed,” he assured my father, “my pole will be length contracted until it’s only
p
1 − (V /c)2 L0 = 35 (100 ft) = 60 ft

long. I’ll be able to fit it completely within your old barn! Look, if you don’t believe it, put me to the test.
Station one of your sons at the front door and the other one at the rear door. Start with both doors closed,
and open each one for just long enough to allow me through. You’ll see. There will be a time when both
doors are shut and my pole is completely enclosed within your barn.”

My father was no dummy. He rubbed his chin and looked puzzled and thoughtful for a minute. “Okay,
you’re on,” he told the pole vaulter. “I’ll station my boys. But there’s just one thing I don’t understand:
Sure, in the barn’s frame your pole will be length contracted. But in your frame the barn is moving. In
your frame my barn is 60 feet wide and your pole is 100 feet long. How are you going to fit that long pole
of yours into my stubby little barn?”

Now it was the champion’s turn to be puzzled. In fact, he looked frightened and just a little greenish. I
could see the sweat bead up on his forehead, and he lost his confident swagger. He wanted to bail out. My
older brother went up and whispered a few words into the vaulter’s ear. They huddled in quiet conversation
for a few minutes, and then the vaulter regained most of his lost confidence. He carried out the feat flawlessly.

What did my brother tell the vaulter?

“You have to think about time and simultaneity also,” my brother whispered, “not just about length.”

“If I’m going to risk my life,” the vaulter responded, “I’ll need a lot more detail than that.”

“Fair enough,” said my brother, as he drew a figure onto the barnyard dust with a stick, “Here’s a sketch
in the barn’s frame. The two important events are when the front door closes and when the rear door opens.
Between those two events you’re completely enclosed within the barn.”

7
 Barn's frame:
event 1:
front door closes

60 ft

100 ft

event 2:
rear door opens

Now, what are ∆x and ∆t between these two events?”

“That’s easy,” replied the vaulter. “∆x is just the length of the barn, 100 feet, and ∆t is just the time I
need to run 40 feet. Since I’ll be running at V = 45 c, the time required is
40 ft
∆t = 4 = (50 ft)/c.
5c
Now I just need to find the speed of light in feet per second. Let me Google it . . . ”
“No, wait,” interrupted my brother. “Before performing the arithmetic, let’s work with the symbol c.
Maybe we won’t need to actually figure out the value in seconds.”
“Well, if you say so,” replied the vaulter dubiously. “I guess next you’ll want to know the separations
∆x0 and ∆t0 in my frame.”
“Let’s begin by finding only ∆t0 . I think that’s all we’ll need. How can we find it?”
“We use the Lorentz transformation, of course.” The vaulter was now looking more peeved than anxious.
He took my brother’s stick and began tracing equations into the dust.
∆t − V ∆x/c2
∆t0 = p
1 − (V /c)2

8
 (50 ft)/c − 54 (100 ft)/c
= 3
5
(50 ft)/c − (80 ft)/c
= 3
5
−(30 ft)/c
= 3
5
= −(50 ft)/c.

“But that’s . . . what . . . a negative number. What could that mean?”

“It means,” said my brother, “that in the barn’s frame, where ∆t is positive, first the front door closes,
and then the rear door opens. Between those two events, the pole is completely enclosed within the barn.
Whereas in the vaulter’s frame . . . .”
“In my frame,” the vaulter jumped in, “those two events happen in the opposite sequence: first the rear
door opens, and second the front door closes. Can that be true? Can it really be that the pole is never
enclosed within the barn in my frame?”
“Sure it can be true,” assured my brother. “As my kid brother likes to say, ‘Two events that occur in
one sequence in one reference frame might occur in the opposite sequence in a different reference frame.’ ”
Problem 3: The pair of sketches in the barn’s frame on page 8 shows (1) the pole’s butt entering the
front of the barn and (2) the pole’s tip leaving the rear of the barn. Draw a similar pair of sketches showing
these two events in the vaulter’s frame. Describe in words the differences between the situation in the barn’s
frame and in the vaulter’s frame.
Problem 4: Repeat the calculation of this section with a barn and a pole of rest length L0 (instead of
100 feet), and a pole vaulter of speed V (instead of 54 c).

The tossed tomato: Causality and speed limits

We have seen that, given two events, it’s possible that in some frames event #1 comes first, in other frames
event #2 comes first, and in one frame the two events are simultaneous.
But suppose that event #1 causes event #2. (For example: event #1 is “I toss a tomato,” event #2 is
“tomato splatters over the wall.”) In this case you’d certainly think that event #1 has to occur before event
#2 in all frames! Let’s make this assumption and see where it takes us.
Define
∆x
= speed of the causal signal = vs .
∆t
(In our example, the “causal signal” is the tomato.)
Our assumption is ∆t0 > 0 in all reference frames, so, by the Lorentz transformation,
∆t − V ∆x/c2
∆t0 = p >0 for all frames.
1 − (V /c)2

9
Thus

∆t > V ∆x/c2
1 > V vs /c2
c2 > V vs

If the causality assumption is true, then this holds for all frames, and for all causal signals.

Suppose the causal signal is light, so that vs = c. Then

c > V,

that is, all reference frames travel at less than the speed of light.

Pick one such reference frame moving just slower than c: that is V = c − , where  can be as small as
you wish as long as it’s greater than zero. Then for any causal signal

c2 > V vs
c2 > (c − )vs
 
c > 1− vs
c
Let  → 0 to find
c ≥ vs ,
that is, any causal signal travels at less than or equal to the speed of light.

Has this prediction been verified? Well, sending information is not just fun to speculate about, it’s a
multi-billion dollar industry. The telecommunications industry, the television industry, the Internet industry,
the computer industry, the courier industry (and, from the past, the Pony Express): all, ultimately, are about
sending information from place to place as quickly as possible. These industries work tirelessly to find ways
of sending information quickly, but none has figured out a way of sending it faster than the speed of light.

Anything else? At the CERN laboratory near Geneva, Switzerland, scientists have figured out how to
push a single electron so hard and so many times that, if common-sense notions of space and time were
correct, the electron would be traveling 904 times the speed of light. But in fact no electron at CERN has
traveled as fast as or faster than light. (The maximum speed achieved so far is 0.999 999 999 997 c.)

So despite the enormous monetary and scientific rewards that would be showered upon anyone sending a
causal signal faster than light, no one has ever been able to do so: good evidence in support of both relativity
and our causality assumption.

Rigidity, straightness, and strength

Here is a proposal for sending a signal faster than the speed of light. . . in fact, for sending it instantaneously:
Push the left end of a rod. . . the right end moves at the same time! Well, not quite. When you push the left

10
end of the rod, you move the first atom in a long chain of atoms that makes up the rod. A short time later,
the first atom pushes the second, then the second pushes the third, and so forth. This push moves down the
rod and reaches the end at a speed that is very fast by human standards (for a steel rod it moves at about
3 miles/second), so we don’t notice it. But the speed is very slow compared to the speed of light. There is
no such thing as a perfectly rigid rod.
Ivan mounts a straight rod horizontally on three pegs, and he fits the base of each peg with a firecracker
that can cause it to crumble into bits. He arranges for the three pegs to crumble simultaneously, and at
that same instant the rod begins to fall down. The rod is always straight and always horizontal. Veronica
moves to the right relative to Ivan. In Veronica’s reference frame these three events are not simultaneous:
First the right peg crumbles and the right end of the rod begins to fall, second the middle peg crumbles and
the middle of the rod begins to fall, third the left peg crumbles and the left end of the rod begins to fall.
Between the first and second events the rod must be curved in Veronica’s reference frame. A rod that is
straight in one reference frame may be curved in another.
A bicycle wheel is set up on a rack and spun very quickly. The pieces of the wheel are moving in the
direction that they are pointing, so they are length contracted. But the spokes are moving in a direction
perpendicular to the direction that they are pointing, so they are not length contracted. How can wheel hold
together with a contracted circumference and a non-contracted radius? The answer is that it can’t. When a
wheel rotates fast it breaks apart and flies into pieces. A wheel made of a weak material like wood will break
apart at rather slow speeds; a wheel made of a strong material like steel will break apart at higher speeds;
a wheel made of a very strong material like diamond will break up at still higher speeds; but a wheel made
up of any material will break apart a speeds much lower than those where the relativistic effect becomes
noticeable. There is no infinitely strong material.
The relativistic limits on the rigidity and strength of materials are can be worked out quantitatively, and
they are extreme. All known materials are much less rigid and much less strong than the limits allow.

Interval

In the problems I have asked you to prove that for any two reference frames, regardless of their relative
velocity V ,
(c∆t)2 − (∆x)2 − (∆y)2 − (∆z)2 = (c∆t0 )2 − (∆x0 )2 − (∆y 0 )2 − (∆z 0 )2 .
This quantity is called “interval” or, by those who like to sound sexy, “spacetime interval”. The problem
solution is nothing but algebra: I did it by starting with the right-hand side, plugging in the Lorentz
transformation equations, and chugging away until I got to the left-hand side. The dramatic moment in the
proof came at the final step: every intermediate step involved the relative velocity V , but in the final step
the expressions involving V canceled out numerator and denominator, resulting in the left-hand expression
independent of V . Here I ask, not about the mechanics of the algebraic chugging, nor about the drama of
the final step, but about the physical significance of this result.
Here’s an analogy involving not relativistic spacetime, but ordinary three-dimensional space. In fact, to
make it easier to draw, I’ll use two-dimensional space. These diagrams show two points in space.

11
 y
2 y' 2

1 1
x'
x

The two diagrams show the same two points, but each shows a different choice of coordinate axes: On the
left, the two points are separated by ∆x and ∆y. On the right, the two points are separated by different
coordinates ∆x0 and ∆y 0 . These are the same two points, with the same separation; the different coordinates
reflect the different coordinate axes, not any difference in the physical points.

Whether we call a tree by the English name “tree” or by the German name “baum” doesn’t make any
difference to the tree; it’s the same tree regardless of name used. The name merely reflects the human
convention of language used, nothing intrinsic to the tree itself. Similarly, whether we call the separation
between two points by the coordinates ∆x, ∆y or by the coordinates ∆x0 , ∆y 0 doesn’t make any difference
to the separation; it’s the same separation regardless of coordinates used. The coordinates merely reflect the
human choice of axes used, nothing intrinsic to the separation itself. This is not to belittle either language
or coordinates: both are essential to human communication. It is to point out that their role is neither more
nor less than human communication.

If you enjoy trigonometry, you will enjoy showing that, if coordinate axis x0 is rotated by angle θ relative
to coordinate axis x, then the coordinates are related through

∆x0 = + cos θ ∆x + sin θ ∆y

∆y 0 = − sin θ ∆x + cos θ ∆y.

Starting from this, you could prove that

(∆x)2 + (∆y)2 = (∆x0 )2 + (∆y 0 )2 .

Your proof would have the same character as the proof that interval is the same in both frames: There’s an
algebraic chug, each step of which involves the angle θ, and then a dramatic surprise at the last step when
several θs cancel out and you uncover the above θ-independent result.

A quantity that is the same regardless of reference frame is called an “invariant” — it’s a quantity that
doesn’t vary from from frame to the next. Invariants are significant because we believe that physical effects
depend only upon the physical situation, and not upon the human choice of coordinate system. For example,
if the two points of the figure were occupied by spheres of mass m1 and m2 , you know that the gravitational
force between them would have magnitude
m1 m2
G ,
∆x2 + ∆y 2

12
which is invariant. If I suggested to you that the gravitational force between the two points were instead
given by
m1 m2
G ,
∆x2 + 7∆y 2
then you would laugh in my face. If this formula were true, then the gravitational force would depend
upon the human choice of coordinate system — which is just as absurd as suggesting that the force depends
upon whether it is described using the English word “gravity” or the German word “schwerkraft”. The
time-dependent interaction between two events in relativistic mechanics is more intricate than the static
interaction between two points in space, but the same idea applies: the interaction has to depend upon the
invariant interval, not upon any frame-dependent quantity.

In the case of separation of two points in space the significance of the invariant

(∆x)2 + (∆y)2

is perfectly clear just from appearance: it’s the square of the distance between the two points. In the case
of separation of two events in spacetime the significance of the invariant interval is not clear from visual
inspection — at least not to me. (The famous book Gravitation by Charles Misner, Kip Thorne, and John
A. Wheeler attempts to make this straightforward on pages 11 and 58, but these pages have never been clear
to me.) But it plays the same role in deciding which forms of interaction are physically permissible.

Distance in space is easy to pictorialize and interpret — it’s just common sense.

Interval in spacetime is difficult to pictorialize and interpret — but you shouldn’t expect it to be, because
nothing in relativity is common sense.

Yet it is our task as scientists to develop such pictures and interpretations. The following four problems
begin this development.

Problem 5. A burst of light is created in a flashlight, travels in a straight line, and is absorbed at a wall
on the opposite side of the room 10 meters away. What is the interval between creation and absorption?

Problem 6. A burst of light is created in a supernova, travels in a straight line, and is absorbed at a
planet on the opposite side of the galaxy 100,000 light-years away. What is the interval between creation
and absorption?

Problem 7. A train departs from New York and arrives in Chicago, 713 miles away, 19 hours later. Is
the interval between departure and arrival positive, zero, or negative? (Such a separation — in which the
“space parts” (∆x)2 + (∆y)2 + (∆z)2 are smaller than the “time parts” (c∆t)2 — is called “timelike”. If
two events are separated by a timelike interval, there is some reference frame [in this example the train’s
reference frame] in which those two events happen at the same location. The first event will precede the
second in all reference frames, and the first event might or might not cause the second.)

Problem 8. In the “pole in the barn” paradox, call “front door closes” event number 1 and “rear
door opens” event number 2. Is the interval between these two events positive, zero, or negative? (Such a
separation — in which the “space parts” (∆x)2 + (∆y)2 + (∆z)2 are larger than the “time parts” (c∆t)2

13
— is called “spacelike”. If two events are separated by a spacelike interval, there is some reference frame in
which those two events happen at the same time. The sequence of the two events might differ in different
reference frames, and one event cannot cause the other.)

Acknowledgment: This section grew out of questions from Rainie Heck, Oberlin College class of 2020.

The twin paradox

Sirius is the brightest star in the night sky, a brilliant blue-white star located just below the constellation
Orion. Any astronomy textbook will tell you that Sirius is located eight light-years away from Earth. This
means that (in the Earth’s frame) it takes eight years for light, traveling at 186 000 miles/second, to fly from
Sirius to Earth. We won’t write down this distance in miles or in meters. Instead, we’ll say instead that the
distance to Sirius is (8 yr)c.
4
Veronica decides to take a trip from Earth to Sirius and back again, traveling at V = 5 c. Veronica’s
friend Ivan chooses to remain at home on Earth.

Let’s get two issues out of the way at the very start: First, the Earth and Sirius move relative to each
other (if nothing else, due to the Earth’s orbit around the Sun) but this relative motion is so slow relative to
4
5 c that we can safely ignore it, and consider Sirius to be at rest in the Earth’s frame. Second, this relative
motion means that the distance from Earth to Sirius changes with time, but these distance changes are so
much smaller than 8 light-years that we can safely ignore them, too.

The high points of Veronica’s journey, as observed from the Earth’s frame, are shown on page 16. (The
rectangular clocks display time measured in years.) You should be able to calculate all of these clock readings
yourself: the only principles employed are the definition speed = distance/time and time dilation.

One more issue requires attention here. Veronica doesn’t just step into her space ship and then instantly
move at 54 c any more than you step into your car and then instantly move at 60 miles/hour. It takes some
time (a “period of acceleration”) for Veronica to get up to her cruising speed. Exactly how much time it
takes will depend on the kind of space ship Veronica uses, but let’s say it’s a week: it can’t be instant,
but the acceleration period can be small compared to the many years of total travel time. Similarly for the
turnaround at Sirius: let’s say it takes two weeks for her to slow down, turn around, and then get up to
cruising speed for the return leg of journey. But since the clock readings shown in the figure are accurate
only to the nearest tenth of a year anyway, these acceleration periods can safely be ignored.

The upshot is that at the end of the trip Ivan’s clock has ticked off 20 years while Veronica’s clock has
ticked off 12 years. This applies not only to wristwatches, but also to biological clocks. Ivan will have aged
20 years and Veronica will have aged 12 years, so there will be more wrinkles on his face than on her face.
Ivan explains this by saying that Veronica’s moving clock ticks slowly. How does Veronica explain it?

We examine the outbound leg of the journey in Veronica’s frame. As usual, changing from the Earth’s
frame to Veronica’s frame involves four differences:

14
 1. Instead of Veronica moving right, the Earth and Sirius move left.
2. The Earth and Sirius are closer (length contraction).
3. Clocks on the Earth and Sirius are not synchronized (relativity of synchronization).
4. The Earth and Sirius clocks tick slowly (time dilation).

I’m not going to detail how I arrived at the numbers on line A at page 17. You should do this for yourself.
Be sure to notice the qualitative character of these results: The distance from Earth to Sirius is shorter in
Veronica’s frame. The Sirius clock is to the rear, so it is set ahead.

How much time elapses before Sirius comes to meet Veronica? This is just
distance traveled (4.8 yr)c
time elapsed = = 4 = 6 yr,
speed 5c

so six years elapse and Veronica’s clock ticks off six years. The Earth and Sirius clocks tick slowly, of course,
so they tick off only 35 (6 yr) = 3.6 yr. These results are reflected in line B at page 17.

The next step, as described from the Earth’s frame, is that Veronica slows down, turns around, and heads
back to Earth. She’s slipping out of one inertial frame, and into another. But from Veronica’s point of view,
the difference is that the Earth and Sirius are no longer moving left. . . instead they’re moving right. On the
outbound leg of the journey Earth is to the front of Sirius, so its clock is set behind by 6.4 years. On the
return leg Earth is to the rear of Sirius, so its clock is set ahead by 6.4 years. Compare closely lines B and
C at page 17. During the turnaround, which requires two weeks for Veronica, the Earth clock jumps ahead
by 12.8 years! This is the relativity of synchronization in action.

The return leg of the journey is like the outbound leg: Veronica’s clock ticks off 6 years and the Earth
and Sirius clocks tick off 3.6 years. Veronica finds that when she returns home, her clocks have ticked off
12 years while Ivan’s have ticked off 20 years.

Ivan explains this difference by saying that Veronica’s clock ticks slowly. Veronica explains it by saying
that Ivan’s clock ticks slowly, except that during the turnaround (when it changed from front clock to rear
clock) it ticked very rapidly indeed.

This effect is called the “twin paradox” because if two twins go separate ways — one staying at home and
the other traveling at high speed — then when they come back together the traveling twin will be younger.

This is the end of the essay, but not the end of relativity.

I have spent four decades pondering and calculating and experimenting with space and time, deepening
and broadening my understanding of relativity, yet still there are facets I find puzzling. For the first decade
or so I found this disorienting and depressing, but I’ve grown to appreciate it: It would be sad indeed if I
understood relativity so thoroughly that it would never again surprise or delight me.

If you’re as lucky as I am, then you too will be surprised and delighted for the rest of your life.

15
Earth's frame

(8.0 yr)c

A. depart Earth: 00.0 00.0

4
5
c
00.0

B. arrive Sirius: 10.0 10.0

4
5
c
06.0

C. depart Sirius: 10.0 10.0

4
5
c
06.0

D. arrive Earth: 20.0 20.0

4
5
c
12.0

Veronica’s voyage to Sirius and back, from the Earth’s reference frame.

16
Veronica's frames

(4.8 yr)c

4 4
5
c 5
c
A. Earth departs: 00.0 06.4

00.0

4 4
5
c 5
c
B. Sirius arrives: 03.6 10.0

06.0

4 4
5
c 5
c
C. Sirius departs: 16.4 10.0

06.0

4 4
5
c 5
c
D. Earth arrives: 20.0 13.6

12.0

Sirius’s voyage to Veronica and back, from Veronica’s two reference frames.

17