SIMPLY EINSTEIN

Relativity Demystified

• • •

RICHARD WOLFSON

W • W • NORTON & COM PANY

NEW YORK • LONDON

DEDICATION
• • •

For Irving Wolfson

and Leonard Swift
CONTENTS
• • •

Cover

Title Page

Dedication

Preface

1 The Self-Creating Universe and Other Absurdities

2 Tennis, Tea, and Time Travel

3 Moving Heaven and Earth

4 Let There Be Light

5 Ether Dreams

6 Crisis in Physics

7 Einstein to the Rescue

8 Stretching Time

9 Star Trips and Squeezed Space

10 The Same Time?

11 Past, Present, Future, and . . . Elsewhere

12 Faster than Light?

13 Is Everything Relative?

14 A Problem of Gravity
15 Into the Black Hole

16 Einstein’s Universe

Appendix: Time Dilation

Glossary

Further Readings

Index

Copyright
PREFACE
• • •

Have you ever heard it said of a difficult idea that “it would take an Einstein
to understand this”? What could be more incomprehensible to us non-
Einsteins than Albert Einstein’s own work, the theory of relativity?
But relativity is comprehensible, and not just to scientists. At the heart of
relativity is an extraordinarily simple idea—so simple that a single English
sentence suffices to state it all. Some consequences of that statement are
disturbing because they violate our deeply held, commonsense notions about
the world. Yet those consequences flow inexorably from a single principle so
simple and obvious that it will take me just a few pages to convince you of
its truth.
This book’s title, Simply Einstein, reflects the fact that the basic ideas of
Einstein’s relativity are accessible to nonscientists and make eminent sense.
Even relativity’s startling implications about the nature of space, time, and
matter follow so directly from those basic ideas that they, too, become not
only comprehensible but also logically inevitable.
Relativity is behind many of the hot topics at the frontiers of modern
physics, astrophysics, and cosmology—topics ranging from black holes to
the ultimate fate of the Universe to the prospects for time travel. I’ll touch on
these topics here, and you’ll see how they flow from the essential ideas of
relativity. But my main purpose is not to explore the latest frontiers of
physics. There are plenty of good books on those topics, and I’ve included
some in the Further Readings. Rather, this is a book that aims to give you, its
reader, a clear understanding of just what it was that Einstein said about the
ultimate nature of physical reality. To help you get there, we’ll be exploring
together the history of ideas that culminated in Einstein’s simple but
remarkable vision. Then you’ll see how that vision alters your commonsense
notions of space and time in ways that would let you travel a thousand years
into the future in just a few short hours. You’ll come to a new understanding
of “past” and “future” that might surprise historians, and you’ll begin to feel
at home in the four-dimensional universe of relativistic spacetime. Along the
way I’ll anticipate your frequent questions: Why can’t anything go faster than
light? Will I really age more slowly, or is this just something that happens to
physicists’ clocks? Can I go backward in time? What does E = mc2 really
mean? Finally, in the end, we’ll return to some of those contemporary hot
topics that show just how prescient was Einstein’s visionary insight.
You don’t need to do math to grasp the essence of Einstein’s relativity,
and you don’t need math to understand this book. Occasional numbers can
help make some points more concrete, and I’ll use them sparingly. What’s
important here are the big ideas—and they’re all expressed in words. Grasp
those ideas, and you know what Einstein’s relativity is all about. Enjoy!
CHAPTER 1

THE SELF-CREATING UNIVERSE AND OTHER

ABSURDITIES
• • •

Could the Universe have created itself? What an absurd idea! Did the
Universe even have a beginning? That question, too, has an absurd ring. If
there was a beginning, what came before? Wasn’t that part of the Universe
too? Or has the Universe always existed, begging the question of its own
origin?
Whatever the answers to these questions, modern astrophysics makes one
thing clear: our Universe hasn’t existed forever unchanged. Rather, it’s
evolved from an earlier state of extreme temperature and density. Some 14
billion years ago, all the “stuff” that makes up ourselves, our planet Earth,
and all the stars and galaxies was crammed into a volume far smaller than a
single hydrogen atom or even the tiny proton at its core. The expansion of that
extreme state is the Big Bang that describes the Universe’s subsequent
evolution and ultimately accounts for the origin of stars, galaxies, planets,
and intelligent life.
What came before the Big Bang? What created that early, extreme state?
We’re back to the primordial question: Did the Universe have a beginning, or
has it always existed—albeit an existence marked by evolutionary change?
To some cosmologists—scientists who concern themselves with the
origin and evolution of the Universe—the start of the Big Bang marks the
start of time itself. For them, it makes no sense to ask what came before
because the concept of “before” is meaningless if there’s no such thing as
time. Others have envisioned an ever existing Universe that undergoes a
series of oscillations. Each begins with a Big Bang and subsequent expansion
—the phase we’re now in—then eventually contracts toward a Big Crunch of
extreme density and temperature that starts another cycle.
In 1998 Princeton physicist J. Richard Gott and his student Li-Xin Li
published a novel answer to the ultimate question of the Universe’s origin.
Their paper, “Can the Universe Create Itself?,” shows how the laws of
physics may allow a time loop, in which time goes round and round in a
circlelike structure rather than advancing inexorably into a never-before-
experienced future. Like Bill Pullman’s character in the film Groundhog
Day, an occupant of the time loop might go to bed at night and wake up on the
morning of the day before! The “new” day would unfold, night would come,
and again the morning would bring the already familiar day. In this loop, time
advances circularly into a future that is a recycling through past events.
There’s no earliest event, any more than any point on a circle can be called
the beginning of the circle or any point on Earth’s surface is the place where
our planet “starts.”
Gott’s time loop doesn’t sound like a description of our Universe, but
hold on—there’s more. Time, in Gott’s theory, can branch, providing
different paths to different futures. Gott envisions a universe whose earliest
epoch includes a time loop. Every point on the loop both precedes and
follows every other point. There is no beginning instant, because one can
always trace time further back round the loop. But there’s a branch out of the
time loop, a branch into a more normal realm of time that advances, without
repetition, to a future of never-before events. That’s the kind of time we
know, with an as-yet-unknown, yet-to-occur future. Figure 1.1 depicts Gott’s
time-loop, multibranched universe.
It’s the branching that reconciles Gott’s time loop with the more ordinary
time we experience today. The time loop unambiguously precedes our
present, and in that sense it’s closer to the beginning. But trace time
backward, through the branch and onto the loop. You can keep tracing back
but you’ll never find a beginning. Instead, events repeat as your historical
exploration circles backward around the loop. There’s no one event that
marks the creation of this universe. Every event on the time loop precedes
every other event, and in that sense Gott’s universe creates itself. Absurd!
Fig. 1.1 Gott’s time-loop universe. Arrows rep resent multip le directions of time, including the circular time loop at the
beginning, in which time goes round and round in an ever rep eating sequence. Circular cross sections of the “trump ets”
rep resent p osition in a single sp atial dimension, and each trump et is its own universe. Is this absurd? M ay be not, say s the
theory of relativity. (Adap ted with p ermission from J. Richard Gott and Li-Xin Li, “Can the Universe Create Itself?” Physical
Review D 58 (1998), p . 3501.)

How Many Universes Did You Order?

Surely, “The Universe” encompasses all that there is. That’s the root meaning
of the word, as in “universal.” But not according to Stanford University
cosmologist Andrei Linde. For the Russian-born Linde, our Universe is but
one small branch of a possibly infinite Multiverse. What we think of as the
Big Bang origin and evolution of the Universe is, to Linde, simply the
“budding” and subsequent expansion of a new branch from a pre-existing
cosmos. That branch is our Universe. Other branches are different universes,
each of which has had its own big bang and its own evolutionary scenario.
Remarkably, each universe may even have its own laws of physics. The
budding that produces a new universe may result in mutations from the laws
that govern the parent branch. Together, all these interconnected universes
form the Multiverse or, in Linde’s more dynamic phrasing, the “self-
replicating inflationary universe.” Our own Universe may someday spawn
new buds that become entire universes; in fact, it may already have done so.
It might not even take much effort to initiate such a bud. Cosmologist Alan
Guth of MIT has suggested that with an ounce of material, crushed to high
enough density, you might start a new universe right in your own garage!
Perhaps we and our whole Universe are just the results of someone’s
experimentation in another branch of the Multiverse.
Linde’s Multiverse provides yet another answer to the question of the
Universe’s origin. Our Universe, according to Linde, clearly had a beginning
in the budding event that was the start of our Big Bang. But that budding
occurred from one branch of a Multiverse that may have existed forever—as
if the structure in Figure 1.1, instead of starting with the time loop, just
continued backward forever in a jumble of budding and branching universes.
That self-replicating Multiverse in some ways resembles a biological
system. It’s forever spawning new buds—“baby universes”—some of which
grow to become full-blown universes like our own, which then produce their
own babies. Others are stillborn, withering to collapse before they’ve had a
chance to evolve complex structure and intelligent life. Universes come and
go, so there are multiple beginnings. Creation isn’t a one-time story. But the
Multiverse persists forever, and, despite the birth and demise of individual
universes, the large-scale picture may remain unchanged for eternity.

Contact!
In the film Contact, based on Carl Sagan’s novel with the same title, actress
Jodi Foster plays the first astrophysicist to detect interstellar signals from an
advanced civilization. The signals convey a message—instructions for
building some sort of machine. Machines are built, against a backdrop of
political and religious intrigue, and eventually Foster’s character boards one
for a ride into the unknown. The machine takes her through a wormhole in
spacetime and deposits her on a distant world where she learns that entire
galaxies are actually cosmic engineering projects. She rides the wormhole
machine back to Earth only to find that no one believes her story because, for
folks on Earth, no time has elapsed while she was ostensibly touring the
cosmos.
Contact looks like science fiction. But many of its key ideas, including
wormholes through space and time, are based in sound physics. In fact,
author Sagan—a scientist himself—consulted colleagues about the validity of
the sci-fi ideas at the basis of Contact. As a result, Sagan’s novel spurred a
flurry of interest in wormholes and in the possibility of time travel. By the
turn of the century, leading researchers had published scores of papers on
these subjects in the most respected physics journals. Some show how
wormholes might connect seemingly distant parts of the Universe that are
actually less than an inch apart in a hidden dimension. Others debate the
mathematical possibility and philosophical implications of time machines
that might let us travel into the past. That’s getting pretty speculative, but
another form of time travel is solidly established. Read on.

Escape to the Future

In the 1970s, scientists sent a very accurate atomic clock on a trip into the
future. How? By flying the clock around Earth on commercial airplanes.
When the clock returned to its starting point, it showed less elapsed time than
a companion clock that hadn’t made the journey. So what? That time
difference means the traveling clock had somehow jumped into the future,
arriving back at its starting place at a time that was further advanced than its
own reading would suggest. For the traveling clock the difference amounted
to some 300 billionths of a second. No big deal!
But we’re convinced that the same idea would work in a more dramatic
context, allowing you to “leapfrog” into the distant future. Here’s the
scenario: You and I, who are about the same age, collaborate in building a
high-tech spaceship capable of traveling at close to the speed of light (about
186,000 miles per second). You board the ship, zoom off to a star in our
galactic neighborhood, and return. As far as you’re concerned, the trip takes
a few days. But on return, you find me some 20 years older than when you
left. You pick up a newspaper, and it’s dated 20 years after your departure.
You look around and see that planet Earth and all of human society have
advanced 20 years during your several-days trip. Somehow you, like the
atomic clock, have jumped into the future. This time the jump isn’t a
negligible fraction of a second but a goodly chunk of a human lifetime. Take
that spaceship further—say, to the center of our Milky Way galaxy—and
when you return, weeks later as far as you’re concerned, you’ll find yourself
some 60,000 years in the future!
Alas, there’s no going back—at least not with this form of time travel. If
you don’t like what you find thousands of years in the future, you can either
put up with it or jump further into the future.

The Universe as Telescope

Much of what we know about the Universe beyond our home planet comes
from telescopes—instruments that collect and analyze light and other forms
of radiation from the cosmos. Before the early twentieth century, telescopes
were too weak to see very clearly beyond our own Milky Way galaxy. That
changed with the completion of the 100-inch Mount Wilson telescope in
1917, and within 10 years observations from Mount Wilson had radically
enlarged humankind’s conception of the Universe—a conception that
included, for the first time, hints that the Universe had a beginning.
Mount Wilson reigned for three decades as the world’s most powerful
telescope, until the 1949 opening of the 200-inch Mount Palomar instrument
quadrupled astronomers’ powers of observation. (Those 100-inch and 200-
inch figures are the diameters of the telescopes’ mirrors; the amount of light
they collect depends on the mirror area, which quadruples for each doubling
of its diameter.) Today’s largest ground-based telescopes have mirror
diameters of 10 meters, or nearly 400 inches. With their huge but flexible
mirrors, prime mountaintop sites, and a host of advanced technological
features, these modern instruments greatly extend our vision of the Universe.
The 1990 launch of the Hubble Space Telescope gave astronomers an
exciting new tool for probing the Universe. Its location above the distorting
effects of Earth’s atmosphere more than compensates for Hubble’s modest
94-inch mirror. Hubble continues to churn out discoveries, ranging from dust
storms on Mars to black holes in neighboring galaxies to objects at the very
edges of space and time.
Today Hubble and all its Earth-based companions have been bested—
reduced to the role of eyepieces for a “telescope” vastly greater in scale.
This “telescope” consists of a cluster of some 10,000 galaxies, located 2
billion light-years from Earth, that acts as a vast cosmic lens. The gravity of
this huge assemblage of matter bends light rays and thus concentrates light
from more distant objects, making them visible to the paltry telescopes of our
own making. In 2001, this cosmic telescope led to the discovery of a “baby
galaxy” that formed some 13.4 billion years ago—only 600 million years
after the Big Bang and in a realm of time astronomers call the “Dark Ages”
because most stars had not yet formed. The cosmic telescope provides a 30-
fold enhancement of the light from the baby galaxy; without this effect, the
galaxy would be utterly invisible even to the most advanced human-made
telescopes. By the way, the cosmic lensmaker didn’t make this lens quite
perfect; like other gravitational lenses, the huge galactic cluster that focuses
light from the distant baby galaxy forms distorted and often even multiple
images of objects that lie beyond it. In fact, it was through such multiple
imaging that gravitational lenses were first identified.

A Common Absurdity

Time loops? A time before time? A many-branched Multiverse? Wormholes?

Leapfrogging into the future? Galactic telescopes peering to the edge of
space and time? What do all these seeming absurdities have in common? All
require conceptions of space and time that boggle our common sense. All
require that space and time bend, warp, and distort in ways that have no
counterparts in our everyday experience. All require that we give up our
tenth-grade notions of geometry, with straight lines and perfect triangles.
Straight lines through space, from here to there? Not in a Universe where
cosmic lenses focus multiple images of the same object! Straight lines
through time, from now to then? Not in a Universe with wormholes, time
loops, and high-speed star trips!
The geometry of space and time is not, in fact, the geometry you learned
in tenth grade. It’s a much richer geometry that allows for the curving,
bending, or warping of space and time. That richness, in turn, enables a host
of new phenomena: time loops that bend time back on itself; black holes,
where spacetime curvature becomes infinite; gravitational lenses that create
multiple paths for light from distant sources to reach Earth; high-speed travel
that’s a shortcut to the future. The strange geometry of spacetime is not some
unfathomable mystery, though. It’s described precisely by Einstein’s theory of
relativity—a theory that is as much about the geometry of space and time as it
is about phenomena of physics. In fact, relativity suggests that we think of
geometry as a branch of physics rather than mathematics, setting the
spacetime “stage” on which we and the rest of physical reality strut and fret
our roles in the larger Universe.
So what is this relativity theory that mixes geometry into physical reality
and leads to mind-boggling possibilities like time loops, wormholes, and
cosmic telescopes? Ultimately, the idea behind relativity is a very simple one
—so simple that I can state it in a single sentence. But it’s a very big idea,
too—an idea whose consequences and philosophical implications go far
beyond the confines of physics. In the next chapter I will introduce you to the
simple Principle of Relativity, and I’ll convince you that you already grasp
and even embrace this idea at the heart of relativity. From there we’ll
explore how relativity came to be, what it really means, and why its
consequences fly in the face of common sense. At the end, I’ll return to the
cosmic implications of relativity to show how exotic phenomena like
wormholes, time loops, and gravitational lenses follow from your newfound
understanding of Einstein’s remarkable vision.
CHAPTER 2

TENNIS, TEA, AND TIME TRAVEL

• • •

Picture this: You’re on a cruise ship, sailing straight and steady through calm
waters. You’re playing tennis on the ship’s indoor court. How does the
moving ship affect your game? As the ball approaches your racket, do you
sense the ship’s motion and adjust your swing accordingly? And what about
your opponent, who has her back to the ship’s motion? Does she have to
adjust her play in a way different from you?
The answer to all these questions is an obvious no. Your game on the
cruise ship proceeds exactly as it would on land. Neither you nor your
opponent needs to consider the ship’s motion whatsoever. Neither of you
needs to adjust your playing style to compensate for that motion.
Here’s something simpler than the complex volleying of a tennis match.
Suppose you’re standing on land, with no wind blowing, and you throw a
tennis ball straight up. Up the ball goes; eventually it slows to a momentary
stop, then drops straight back down into your waiting hand. Now try the same
thing in that indoor court on the cruise ship. (I’ve put the court indoors to
eliminate the wind you would experience on an outdoor court because of the
ship’s motion through the air.) Again the ball goes straight up and drops
straight back down. That the ship is moving doesn’t matter in the least.
Tired of tennis, you stroll over to a microwave oven to heat a cup of tea.
Now how do you compensate for the ship’s motion? Do you put the cup at the
sternward end of the oven because the microwaves get “left behind” by the
moving ship? Do you adjust the power level because the microwaves behave
differently from how they would back on shore? Of course not! The oven,
like the tennis ball, behaves in just the same way on the moving ship as it
would on solid ground.
Of course, a cruise ship doesn’t go very fast—maybe 30 miles per hour
or so. So imagine now that you’re in a space colony on Venus, which at that
point in its orbit happens to be moving about 30 miles per second with
respect to Earth. The atmosphere inside the colony is the same as on Earth,
and Venus’s gravity has essentially the same strength as Earth’s. You step into
the colony’s recreation center for a game of tennis. As you start your serve,
do you need to account for the fact that the whole court is moving at 30 miles
per second? How could you possibly compensate for that speed? The
answer, of course, is that you don’t have to. The tennis game on Venus
proceeds just as it would on Earth. Venus’s motion is irrelevant.
Now try the simpler act of throwing the ball straight up. One second later,
is it 30 miles away from you? Of course not! It goes straight up and comes
straight down, just as it would on Earth. And why not? Why should the
existence of Earth, millions of miles away and moving at 30 miles per
second with respect to Venus, have any significant effect on what happens on
Venus? Why should the rules governing the tennis ball be any different on
Venus than they are on Earth?
Tired of Venusian tennis, you stroll over to a microwave oven to heat a
cup of tea. This oven, of course, is moving at 30 miles per second. Do you
worry about that? Of course not! Again, Venus’s motion is irrelevant.
Maybe even 30 miles per second isn’t much. So now imagine you’re a
humanoid being on an Earthlike planet in a distant galaxy moving away from
Earth at 80 percent of the speed of light—about 150,000 miles per second.
(The Hubble Space Telescope routinely observes very distant galaxies with
speeds this high; their motion is part of the overall expansion of the
Universe.) Again, it’s tennis time. As you start your game, do you think about
the fact that you’re moving at 150,000 miles per second away from Earth,
and do you make a hopeless attempt to compensate for that colossal motion?
Of course not! You don’t even know about Earth; the Milky Way galaxy, in
which Sun and Earth are minor players, is but a faint smudge in your
civilization’s most advanced telescopes. Why should Earth have anything to
do with you and how you play tennis?
Again, the game’s over and it’s tea time. The microwaves that heat your
tea move through the oven at the speed of light—and this time the oven itself
is moving at 80 percent of that speed. Do the microwaves get left behind?
Does the microwave oven need a different set of instructions—or a whole
different design—to work on this distant planet because it’s moving at such
high speed? The answer, again, is a firm no. The microwave oven works just
fine, thank you, on this distant planet as it does on Earth.
These three examples of the cruise ship, Venus, and the distant galaxy all
illustrate a point that seems to make perfect sense: The physical “stuff” of the
Universe (tennis balls, atoms, microwaves—indeed, all matter and energy)
works the same way everywhere. Put another way, the laws of physics—the
rules that are somehow written into the structure of the Universe and that tell
things how to behave—are the same everywhere. There’s no special,
preferred place where the laws of physics are correct while they need
modifying everywhere else.
This idea that no place is special is hardly new. In 1543, Copernicus’s
On the Revolutions of the Heavens first challenged the ancient view of Earth
as the center of the Universe. Copernicus’s Sun-centered theory was a major
shift not only for science but also for philosophy and religion (which is why
the established church fought Copernicus’s ideas). Since Copernicus,
discoveries in science have only reinforced the view that there’s nothing
special about Earth in the cosmic scheme of things. We, its inhabitants, may
love our planet and consider it a special place—but that doesn’t mean Earth
gets special privileges where the laws of physics are concerned.
Actually, “place” isn’t quite the right word here. While it’s true that the
laws of physics apply everywhere, more important in my tennis examples is
that they apply in different states of motion. A tennis match and a microwave
oven—both governed by the laws of physics—work the same way on Earth
and on a hypothetical planet moving away from Earth at 80 percent of the
speed of light. The fact of the planet’s motion is simply irrelevant. If it
weren’t irrelevant, there would be something mighty special about Earth, in
that everyone else in the Universe would have to reference Earth in
describing their own physical situations.
This idea that motion doesn’t matter in some absolute sense goes back to
Galileo and Isaac Newton, who were first to formulate quantitative laws that
describe the way objects move. But the idea found its complete fruition with
Albert Einstein, whose theory of relativity is based on the simple statement
that motion doesn’t matter, that is, that the laws governing physical reality are
just the same on that distant planet hurtling away from Earth as they are right
here on Earth itself. I’ll rephrase that absurdly simple fact: The theory of
relativity is, in its barest essence, just the simple statement that regardless of
one’s state of motion the laws of physics are the same.
If you followed the simple example of the tennis match and the
microwave oven, then you already know and believe the basis of Einstein’s
theory of relativity. At its heart, this really is the essence of relativity: that
you can play tennis and use a microwave oven—or undertake any other
activity involving physical reality—in any state of motion, and you will
always get exactly the same results. Put in more scientific language, the
results of any scientific experiment will be the same for anyone who cares to
undertake that experiment—again, regardless of state of motion. By
“experiment,” I simply mean some activity that probes the behavior of
physical reality. In that sense, playing tennis is an experiment. So is heating
water in a microwave oven. Watching the tennis ball bounce or the water
boil tells you about how the physical world works. A simpler and more
controlled experiment is the third one I introduced: throwing a ball straight
up and observing its subsequent behavior. What relativity says about these
and any other scientific experiments is that otherwise identical experiments
performed by experimenters in different states of motion will give the same
results. By “identical” I mean all relevant physical circumstances are the
same. That’s why I put the tennis court on Venus and on an Earthlike planet in
a distant galaxy; that way, gravity was the same and so, therefore, was the
force with which you had to swing the racket to put the ball in your
opponent’s court. On Jupiter or the Moon your tennis game would go a bit
differently because of the stronger or weaker gravity, but ultimately you
could infer the same underlying physical laws. Stated more directly, anyone
who cares to experiment with physical reality will discover the same
fundamental physical laws, regardless of the experimenter’s motion.
So you already know and believe the essence of relativity. In that sense,
you can close this book and be done with it. Stripped to its essentials, the
theory of relativity is summarized in a single, simple English sentence: The
laws of physics are the same for all. This statement—called the Principle of
Relativity—is so obvious that you need read no further, so obvious that I’ve
probably belabored it too much already.
 But wait! Although the essential idea of relativity is simple,
straightforward, and so eminently believable as to be obvious, the
consequences of that idea are anything but transparent. Nor, at first glance,
are those consequences particularly believable. I gave you some examples in
the preceding chapter; they included such seeming absurdities as time loops,
wormholes, and shortcuts to the future.
Consider the last of these, in which you and I, originally the same age,
find ourselves 20 years apart when you return from a trip to a nearby star. Is
this really possible? On what absurd new principles of physics does it rest?
It is possible and was verified, albeit modestly, by the round-the-world
atomic-clock experiment I described. For high-speed subatomic particles, the
observed effect is even more dramatic than for your hypothetical star trip.
But there’s no bizarre new principle involved—only the principle that the
laws of physics should be the same regardless of one’s state of motion.

Common Sense or Uncommon Sense?

The story of your jump into the future is hard to swallow because it seems to
violate everything your common sense tells you about the nature of time and
space. Well, your common sense is wrong! Without going into the details of
just how it’s wrong (they’ll come out in subsequent chapters) I’ll simply say
that your common sense is wrong because it’s based on a very limited
experience that masks the full richness that the physical Universe has to offer.
No matter how cosmopolitan, no matter how cultured and well read, and no
matter how much of a world traveler you are, it’s a simple fact—no offense
intended—that you’re very limited and provincial in your outlook. Why?
Because you’ve spent all your life on a small, cool chunk of rock called
Earth. Unless you’re an astronaut, a military pilot, or have flown on the
Concorde, the fastest you’ve ever moved relative to things that are
significant to you is about 600 miles per hour in a jet airplane. That
italicized phrase is important, because you do move at much greater speeds
relative to some other things. For example, you move at some 30 percent of
the speed of light relative to the electrons that beam through your TV to
create the picture you see on the screen. But you have no direct experience of
those electrons, so they play no role in establishing your commonsense
notions. And you’re moving at more than 90 percent of the speed of light
relative to the most distant galaxies observed to date—but again, you don’t
experience those galaxies in any significant way, so they play no role in
developing your commonsense ideas. Were you an engineer designing TV
picture tubes or an astrophysicist studying distant galaxies, it would be
obvious to you that something isn’t quite right with your commonsense
notions of time and space.
Think a minute about how those notions developed. As a baby, you
crawled about the house, maybe tumbled down the stairs, and gradually
formed your concept of three-dimensional space. Later you walked, ran,
bicycled, drove, or flew—and all those activities only reinforced your
childhood understanding of space as a kind of absolute stage for the activities
of your life. As a young child on a long car trip, your plaintive “When will
we be there?” showed an embryonic conception of time. That, too, grew into
your adult notion of time as an absolute, universal measure of progress from
past toward future. But again, that sense of an absolute space and time—with
you and your friends occupying the same spatial stage and aging at the same
rate—grew out of your very limited experience on this one planet. As far as
concepts of space and time are concerned, the particularly limiting aspect of
your experience is your rather slow movement in relation to anything
significant to you.
If you, as a baby, had crawled about at nearly the speed of light—again,
relative to important things like your parents or the household furniture—then
you would have no need of this book, and the world would not have needed
Einstein. Relativity and its consequences would be obvious to you and
having a friend jump into the future would be entirely consistent with your
commonsense notions of time and space.
If you accept the point of this chapter—that the laws describing physical
reality should be the same for everyone—then the basic principle behind
relativity already makes sense to you. But consequences that follow logically
from that principle seem to make no sense—like that 20-year gap between
our ages following your star trip. In the remainder of this book I want to
show you just why we’re so sure of the Principle of Relativity and how that
principle leads inexorably to conclusions that seem to fly in the face of
common sense. At the other end, I hope you’ll come to accept those
conclusions and the phenomena they entail as manifestations of a Universe
whose richness is far greater than your everyday experience would suggest.
CHAPTER 3

MOVING HEAVEN AND EARTH

• • •

I’ve just convinced you that the laws governing physical reality should be the
same on Earth, on Venus, and on a planet in a distant galaxy moving away
from Earth at nearly the speed of light. My task was easy because you
already knew that Earth is not the center of the Universe and you accept that
astrophysicists routinely observe distant galaxies moving rapidly away from
us. If the laws of physics apply throughout the Universe, surely they should
work as well in those distant galaxies as they do here—especially given that
neither we nor denizens of distant galaxies have any claim to centrality in the
grand scheme of things. So you came away from the preceding chapter
realizing that you already understand Einstein’s Principle of Relativity—the
simple statement that the laws of physics don’t depend on your state of
motion.
After showing that you already believe the basic premise of relativity, I
reviewed the strange tale of the star trip that could leapfrog you into the
future. That was disturbing because it violated what your common sense tells
you about the nature of time, and it was but a glimpse of other disturbing
things to come. So, although the idea behind relativity is simple, its
consequences are not. I could plunge further into those consequences right
now, but you’ll find it more convincing if we first explore how the concept of
relativity arose and why it required a genius of Einstein’s stature to assert
what is, in its essence, such a simple idea.
In the next few chapters I’ll back up and give a brief historical overview
of the evolution of our understanding of physical reality. I’ll show how the
idea of relativity became inevitable, but not until the early years of the
twentieth century. As you contemplate this history, remember that you have
the advantage of knowing a lot more than even the most advanced scientists
of earlier times. Pre-Copernican scientists lacked your broader perspective
on a Universe in which Earth’s place is not special. It was only around 1920,
when Einstein’s relativity was already becoming widely accepted, that
astronomers recognized the existence of other galaxies. And not until the last
quarter of the twentieth century did instruments like the Hubble Space
Telescope make possible the discovery of distant galaxies moving relative to
us at speeds approaching that of light. The notion of a tennis match in such a
galaxy—a notion that helped convince you of the obviousness of the
relativity principle—would have had little credibility much before your time.

A Matter of Motion
The history of physical science is intimately connected with our
understanding of the nature of motion. If you ever took a high-school physics
course, you studied motion—a subject you may not have found particularly
exciting. But the study of motion is profound, for several reasons. First,
motion is the source of all change. Imagine a world without motion: Earth
stops rotating, so it’s perpetual daytime. Earth stops revolving around the
Sun, so it’s always the same season. Your body can’t move, so you’re stuck
forever in one spot. Atoms cease moving, so there’s no chemistry—no
release of energy, no change in the substances of the world. Nothing evolves,
transforms, mutates, develops, or otherwise changes. And there are no
molecules jumping across the synapses of your brain, so there’s no thought.
Absent motion, everything stops. Period.
Conversely, if motion does exist (as it obviously does), then
understanding motion will help you understand night and day, the seasons, the
chemical reactions that result from the motion of atoms and the electrons
within them, and even the functioning of your brain as it’s based in the motion
of molecules. Understanding motion at the most fundamental level goes a
long way toward explaining the behavior of matter.
Furthermore, you’ve already seen hints that relativity is going to do
strange things to your commonsense concepts of time and space. Those
concepts are themselves closely associated with the idea of motion. What
does it mean to move? It means getting from one place to another, and doing
so in some time. Whatever else motion means, it involves passing through
time and through space. So motion holds the key to understanding time and
space.

Two Questions about Motion

What causes motion? That seems an obvious question and one that probably
suggests some obvious answers. Push a chair, and it moves across the floor.
Stop pushing, and it stops. Push your car, by stepping on the gas and letting
the engine and tires do the actual pushing, and your car goes. Take your foot
off the gas, and the car soon stops. Drop a ball, and Earth’s gravity pulls on
it, causing it to move. Pushes and pulls seem to be the causes of motion. The
physics word for pushes and pulls is force, a word that I’ll use
interchangeably with push or pull. Some forces are obvious, like the direct
push of my hands on that chair; others, like the pull of Earth’s gravity or the
repulsion of two magnets, are mysteriously invisible.
To the ancient Greeks, especially Aristotle (384–322 BCE) and his
followers, the idea that force causes motion was obvious. An ox pulled an
oxcart, making it move. Stop the ox and the cart stops. Other examples took a
little more explaining. An arrow, for example, whizzes through the air
without an obvious push. Aristotelians explained this motion by saying that
air rushes from ahead of the arrow to behind it, exerting the force that keeps
it moving. So for Aristotelians the answer to our question was obvious and
similar to the answer you would probably give: force causes motion.
There’s a more subtle question we can ask about motion: Is there a
natural state of motion? By a natural state, I mean a state of motion that
requires no explanation—a state that an object naturally assumes unless
something is explicitly done to it, like pushing or pulling it. For the
Aristotelians, the answer to this second question is implicit in the answer to
the first. Since it takes a force—a push or pull—to sustain motion, the natural
state must be the state of rest. If you see something at rest, that’s expected; it
requires no explanation. But if you see something moving, you want an
explanation, an answer to the question, What’s making it move?
 For the ancient Greeks our questions about motion admit not one but two
answers. That’s because the ancients distinguished two distinct realms of
physical reality, the terrestrial and the celestial. It’s in the terrestrial realm
that the natural state is to be at rest—meaning, given the terrestrial context,
being at rest with respect to Earth. In fact, the Greeks’ idea of motion’s
natural state incorporates also a simple explanation of gravity: the natural
state for objects in the terrestrial realm is to be at rest as close as possible to
the center of Earth, which itself was the center of the Universe.
The heavens, in contrast to the base, ordinary realm of Earth, are the
abode of the gods and the realm of perfection. Here the stars and planets
move naturally in the most perfect of paths, namely, circles centered on
Earth. So the natural state of motion in the celestial realm is circular motion
—a state that is so obvious as to require no further commentary.
There’s a slight problem with this picture of simple circular motion
throughout the heavenly realm, in that it can’t quite explain the subtle motions
of the planets. Observed night after night, the planets seem not to go
uniformly in the same direction as would be the case if they simply circled
Earth, but on occasion they backtrack in what astronomers call retrograde
motion. In the second century CE, the astronomer Ptolemy solved this
problem by positing that the planets actually move in smaller circles (called
epicycles) carried around on larger circles that are themselves centered on
Earth. By adjusting the size and speed of the motion on these smaller circles,
Ptolemy was able to match closely the observed planetary motions. Figure
3.1a shows a simplified view of the Ptolemaic scheme of the Universe.
Ptolemy’s epicycles helped preserve two essential philosophical bases of
ancient science: (1) Earth is at the center of the Universe, and (2) perfection
reigns in the celestial realm, here in the form of motion in perfect circles.
Fig. 3.1 Simp lified views of (a) Ptolemy ’s Earth-centered universe and (b) Cop ernicus’s heliocentric sy stem. In both sy stems
all motions in the heavens involve p erfect circles, which for the p lanets (black dots) become small circles about larger circles.
Neither diagram is to scale and both omit some details.

Heresy!

Among the most profound advances in all of science is Polish astronomer

Nicolaus Copernicus’s suggestion that the Sun, rather than Earth, is the center
of the Universe, the point about which all other celestial bodies revolve
(Figure 3.1b). Copernicus (1473–1543) first hit on this idea early in the
sixteenth century and expounded it fully with the 1543 publication of his De
revolutionibus orbium coelestium libri vi (“Six Books Concerning the
Revolutions of the Heavenly Orbs”). This new idea—heretical because
established church dogma was rigorously Aristotelian—was the first in a
series of intellectual steps that stripped Earth of its special place in the grand
scheme of things. Einstein’s relativity, as I outlined it in the preceding
chapter, is the ultimate step in that series.
Copernicus’s heliocentric proposal was truly revolutionary, but in other
ways his new model for the Universe clung conservatively to older ideas.
Although he removed Earth from its central place, Copernicus nevertheless
maintained the distinction between terrestrial and celestial realms. For him,
the heavens remained the realm of perfection. Sun, planets, and stars were
perfect bodies, and they moved, appropriately, in perfect circles. Having
Earth and the other planets in circular paths around the Sun helped explain
many astronomical observations that had puzzled the ancients. In particular,
the retrograde motion of the planets and their apparent brightening and
dimming over time both followed naturally from the rather complicated path
a Sun-orbiting planet describes when viewed from a Sun-orbiting Earth.
Copernicus’s placement of Earth among the other planets raised the question
of how the terrestrial realm, with its imperfections and pestilences, could be
part of the heavens. Nevertheless, Copernicus maintained the
terrestrial/celestial distinction and did not change the Aristotelian answer to
the question, Is there a natural state of motion? On Earth, the Copernican
answer remained “at rest” and in the heavens it remained “motion in perfect
circles.”
 Copernicus’s ideas inspired the Danish astronomer Tycho Brahe (1546–
1601) to undertake extensive, regular, and highly accurate observations of
stars and planets. Tycho ultimately rejected both the Ptolemaic and
Copernican views in favor of his own more complicated Earth-centered
model. When Tycho died, he left his prolific set of astronomical data to his
assistant, the German astronomer Johannes Kepler (1571–1630). Working
especially with Tycho’s data for Mars, Kepler conceived a radically simple
way to avoid the complex combinations of circles needed in the models of
Ptolemy, Copernicus, and Brahe. Compounding the Copernican heresy,
Kepler dispensed with the notion that celestial bodies must move in perfect
circles. Instead, he showed, the simplest explanation for planetary
observations was that the planets move in elliptical orbits, with the Sun not
at the center but at a special point called the focus of the ellipse (see Figure
3.2). And instead of incorporeal spirits pushing planets in their circular paths
about a passive Sun, Kepler proposed that the Sun itself exerts the push that
gives the planets their elliptical motion. Furthermore, Kepler succeeded in
formulating quantitative laws that describe how the speed of the planets
varies as they move about the Sun. Although Kepler described planetary
motions accurately, his theory couldn’t explain them. Philosophically, though,
he had taken a big step—freeing the celestial realm from the constraint of
perfectly circular motion.

Fig. 3.2 Kep ler showed that each p lanet moves in an ellip tical orbit, with the Sun at a sp ecial p oint called the focus. The ellip se
shown here is highly exaggerated; most p lanets’ ellip tical orbits are nearly but not quite circular.

Celestial Acne and Other Galilean Heresies

While Kepler struggled to interpret Tycho’s voluminous astronomical data,
his contemporary, the Italian Galileo Galilei (1564–1642), was also reading
“the book of nature.” Galileo insisted on the primacy of experiment and
observation in establishing scientific truth, moving science from what had
been essentially a branch of philosophy to an empirical and precisely
mathematical description of physical reality.
In 1609, Galileo constructed the first astronomical telescope. With this
new tool, he soon found convincing evidence supporting both Copernicus’s
heliocentric Universe and Kepler’s abandonment of celestial perfection.
Turning his telescope on Jupiter, Galileo discovered the four largest of the
many moons that orbit the giant planet. Now known as the Galilean satellites,
these four miniature worlds circle Jupiter with periods of a few days. As you
can verify with binoculars, their position thus changes nightly. Often fewer
than four are visible, as they pass behind or in front of Jupiter itself. To
Galileo, the Jovian system was obviously a miniature version of
Copernicus’s heliocentric solar system, with Jupiter playing the role of the
Sun and its moons, the planets. If Jupiter can host such a system of satellites,
reasoned Galileo, why not the Sun?
Galileo’s support of Copernicus’s heliocentrism strengthened further with
his telescopic observations of Venus. Like Earth’s moon, Venus exhibits
phases that result from our seeing different portions of the planet illuminated
with sunlight. Furthermore, when Venus is full it appears smallest, and when
it’s a narrow crescent it appears largest. That’s not easy to explain with an
Earth-centered model in which Venus remains the same distance from Earth.
But it follows naturally in the heliocentric model, since Venus needs to be on
the other side of the Sun—as far as possible from Earth and thus appearing
smallest—for us to see it fully illuminated.
When Galileo turned his telescope on the Moon, he discovered its
rugged, mountainous terrain—making the Moon a body much more Earthlike
than the purportedly perfect spheres of the celestial realm. Galileo also
became one of the first Europeans to study sunspots—those dark blotches that
we now know as cooler regions of intense magnetism on the solar surface.
(The Chinese had discovered sunspots much earlier and called them
“crows,” but it wasn’t clear that they were actually features on the Sun
itself.) Like the cratered, mountainous Moon, the blemished Sun gave
evidence for imperfection in the celestial realm.
 Alas for Galileo, his scientific efforts on behalf of Copernicus’s ideas
won him no favor with those in power. For Galileo’s staunch advocacy of the
heliocentric heresy, the Inquisition in 1633 sentenced him to life
imprisonment, and he spent the rest of his days under house arrest. Repeated
attempts failed to clear his name and Galileo’s Dialogue Concerning the
Two Chief World Systems, Ptolemaic & Copernican was on the Vatican’s
banned book list until 1835. Not until 1992 did the Pope issue a statement
essentially exonerating Galileo.
Galileo explored not only the heavens but also the terrestrial realm.
Legend has it that he dropped stones of different masses off the Leaning
Tower of Pisa and observed that they hit the ground at the same time.
Whether or not the legend is true, Galileo did argue on logical grounds that
all objects must fall at the same rate. To study gravity further, he rolled balls
down inclines so as to dilute the effect of gravity and thus permit more
detailed observation. Galileo noticed that a ball rolling in a trough-shaped
incline rose to the same height it had started from. On a more gradual incline,
the ball would roll farther horizontally before reaching its starting height.
From this Galileo concluded that the ball would roll forever on a purely
horizontal plane (see Figure 3.3).
Fig. 3.3 This “thought exp eriment” led Galileo to the conclusion that a moving object should continue forever in uniform
motion unless a force acts to change its motion.

Galileo’s conclusion represents a dramatic shift in our understanding of

motion, for it suggests that motion itself needs no cause or explanation.
Specifically, an object moving horizontally in a straight line with constant
speed will continue to do so unless something acts explicitly to slow it
down, speed it up, or change its direction. For Galileo, our two questions
about motion have new answers. What is the natural state of motion? For
Galileo, it’s no longer rest, but motion at constant speed in a straight line.
That state—called uniform motion—requires no explanation. And what
causes motion? For Galileo, that’s no longer the right question. Motion itself
is natural, requiring no cause or explanation—at least if it’s uniform motion,
unchanging in speed or direction. What does require explanation is any
change in motion. And what causes such change? A force, that is, a push or a
pull. So for Galileo, forces don’t cause motion; rather, they cause changes in
motion.
Motivation for Galileo’s work on terrestrial motion came from a problem
inherent in the Copernican scheme. If Earth moves around the Sun, why don’t
we feel Earth’s motion? Why aren’t we left behind or, since we’re not, why
don’t we feel a push or pull to keep us moving along with Earth? Galileo’s
new understanding of motion provides the answer. Since we and everything
else on Earth partake of our planet’s motion, we continue naturally to move
with the Earth—a state of motion that requires no further explanation.
(Earth’s motion isn’t exactly uniform, since the planet is both rotating on its
axis and revolving about the Sun, but this is a minor complication to an
essentially solid argument.) So we don’t feel Earth’s motion because we
share it and are therefore at rest relative to Earth, even though in the
Copernican scheme we and Earth are “really” moving.

The Genius of Cambridge

The year 1642 saw the death of Galileo and the birth of Isaac Newton, whose
genius was to dominate physics until Einstein’s time. After a decidedly rough
childhood, Newton enrolled at Cambridge University in 1661. But in 1665,
just after Newton received his bachelor’s degree, the plague swept across
Europe. The universities closed and students scattered to the countryside.
Newton returned to his native Woolsthorpe for a period of 2 years. But he
was far from idle: during that time Newton developed much of his theory of
light, explored the orbital motions of the Moon and planets, and laid the
groundwork for his theory of gravity.
A popular tale has Newton sitting under an apple tree, being struck on the
head by a falling apple, and thus discovering gravity. But it didn’t take
Newton’s genius to recognize gravity in a falling apple. Our ancestors,
swinging through the trees, already knew instinctively about gravity here on
Earth. If there’s a deeper truth to the apple tale, it’s that Newton, while sitting
under the tree, was contemplating the Moon. And his remarkable stroke of
genius was this: Newton realized that the motion of the apple and the motion
of the Moon are the same motion. Both are “falling” toward Earth, both
under the influence of the same force, namely, Earth’s gravity. Newton coined
the term “gravity” (from the Latin gravitas, or heavy) to describe this force,
and he proposed that gravitation is a truly universal phenomenon. Every
object in the Universe, Newton claimed, attracts every other. The force of
attraction depends on the masses of the objects and the distance between
them. Move two objects farther apart and the gravitational force between
them weakens. Specifically, if you double the separation, the force drops to
one-fourth of its original value; triple the separation, and it drops to one-
ninth. That is, the force drops as the inverse square of the distance between
the two objects. Newton verified this mathematical property by comparing
the acceleration of the falling apple with that of the Moon, some 250,000
miles from Earth and thus subject to much weaker gravity.
If gravity acts between all objects, why don’t everyday objects fall
together in a great clump? The answer is that gravity is an extraordinarily
weak force. Only large accumulations of matter—like planets and stars—are
sufficiently massive to have obvious gravitational effects. For smaller
objects, other forces usually overwhelm gravity. Nevertheless, it’s possible
with sensitive equipment to measure the gravitational force between objects
as small as apple-sized lead spheres. And at the opposite scale, Newtonian
gravitation accounts even for the slow, stately motion of the galaxies
themselves. Such observations of gravity, from laboratory to cosmos, confirm
the remarkable scope of Newton’s ingenious idea.
 Before exploring Newton’s work further, let’s pause to note the radical
philosophical shift inherent in Newton’s recognition of universal gravitation.
Now, for the first time, the same laws describe motion on Earth and in the
heavens. No longer are there two realms with two different sets of physical
laws. Instead, a single physics governs the entire Universe, from everyday
events on Earth to the most distant reaches of the cosmos. With universal
gravitation, Newton has taken another giant step away from the notion that
Earth is somehow special—a step that helps pave the way for relativity.
How is it that both apple and Moon are falling, when the apple is on a
collision course with the ground while the Moon remains forever in its orbit?
Both are falling in the sense that both are moving under the influence of
gravity alone, and both are deviating from the straight-line, constant-speed
motion that is the natural state in the absence of any force. Furthermore, for
both apple and Moon the direction of the deviation is the same—toward
Earth. Thus Earth’s gravity exerts a force on both apple and Moon, and that
force does what Galileo recognized that forces do: it causes change in
motion.
The Moon’s motion differs from the apple’s in its specifics but not in the
fundamental fact that both motions constitute “falling” under the influence of
Earth’s gravity. The difference between the two motions is that the apple
drops, starting from rest, and thus falls straight down. The change in its
motion consists entirely of an increase in speed. The Moon, in contrast, has a
substantial speed in a direction at right angles to “down.” Absent any force,
the Moon would move in a straight line at constant speed. Earth’s gravity
changes that motion, pulling the Moon into the circular orbit that it follows
repeatedly. The Moon is falling toward Earth in the sense that that’s the
direction of its deviation from what would be a straight-line path, but
because of its sideways motion it never gets any closer to Earth.
Using the reasoning we’ve just applied to the Moon and the apple,
Newton became the first to envision the possibility of artificial satellites. He
imagined Earth with a high mountain and considered what would happen to
objects launched horizontally from the mountaintop (Figure 3.4). Give the
object just a little speed, and it falls to Earth, landing near the foot of the
mountain. Give it more speed and it strikes Earth farther from the mountain.
In each case, gravity pulls the object out of the straight-line path it would
follow in the absence of any force. Give the object just the right speed and
the deviation from the straight-line “natural” motion is just enough to follow
Earth’s curvature. At that point the object is in a circular orbit, always falling
toward Earth but never getting any closer. That’s essentially the situation of
the Moon and of many artificial satellites. Throw the object still faster and
its orbit becomes an ellipse—just as Kepler had described for planetary
orbits.
Universal gravitation is only one of Newton’s many contributions to
physics. Building on Galileo’s work, he made precise the idea that the
natural state of motion is uniform motion—straight-line motion at constant
speed—and that forces cause changes in motion. Newton’s first law restates
that point, declaring that an object at rest remains at rest and an object in
uniform motion remains in uniform motion, unless it’s subject to some force.
His second law tells just how much change a given force causes. That change
is characterized by the object’s acceleration, or rate of change of motion.
The acceleration, says Newton’s second law, is larger for a larger force and
smaller for an object of larger mass. If you’ve studied physics before, you’ve
seen the second law expressed in the well-known equation force = mass ×
acceleration. For Newton, acceleration doesn’t just mean speeding up.
Rather, any change in motion—whether in speed, direction, or both—entails
acceleration. The falling apple gains speed as it approaches Earth, and thus
it’s accelerating. Moon’s speed doesn’t change, but the direction of its
motion does because it’s describing a circular orbit. So its motion is
changing, and thus it too is accelerating.
Fig. 3.4 Newton envisioned launching p rojectiles from a high mountain. The faster they ’re launched, the farther they go before
hitting Earth. Fast enough, and a p rojectile “falls” around Earth in a closed orbit. Even in this orbit, the p rojectile is falling in
the sense that it’s deviating from the straight-line p ath (shown) that it would follow in the absence of Earth’s gravity.

Newton’s third law is probably familiar to you in the statement “for

every action there is an equal and opposite reaction.” Actually, that language
obscures the third law’s simple physical meaning. What the law really says
is that if I push on something, then that thing pushes back on me with an equal
force—equal, that is, in strength, but opposite in direction. More generally,
forces always come in pairs: if A pushes on B, then B pushes back on A.
Applied to gravity, the third law means that when Earth pulls on the falling
apple, the apple pulls with equal force back on Earth. Because the apple’s
mass is small, Newton’s second law shows that the apple gets a hefty
acceleration. In contrast, Earth is enormously massive, so its acceleration is
negligible. But in principle Earth does accelerate slightly toward the apple.
Both forces—Earth on apple and apple on Earth—are needed for a full and
consistent description of motion in Newton’s physics.
Newton’s work provided the basis for a full description of phenomena
involving motion. Collectively, Newton’s laws and all that follows from
them compose the branch of physics called mechanics.

A Theory of Everything?
Newton’s three laws of motion, coupled with his law of universal
gravitation, seemed capable of explaining essentially all physical
phenomena. With Newtonian mechanics as a foundation, physicists could
understand not only motion in the cosmos but also such seemingly unrelated
phenomena as the behavior of gases under conditions of changing
temperature and pressure. Although at first glance gas behavior seems to
have little to do with Newton’s laws, in fact it follows directly by applying
those laws to the motions of the individual molecules constituting the gas. So
great was the triumph of Newtonian physics that science held the hope of
explaining all physical phenomena ultimately in terms of the motions of
particles obeying Newton’s laws.

Galilean Relativity
At the heart of Galileo’s and Newton’s new understanding of motion is the
idea of uniform motion as a natural state, needing no further explanation. It
was this idea that helped Galileo accept Copernicus’s moving Earth.
Because everything on Earth shares a common motion and continues naturally
in that state, Earth’s motion is not obvious to us. Indeed, both Galileo and
Newton knew that the laws of motion didn’t favor a particular place or state
of motion. Galileo reasoned, for example, that a rock dropped from a ship’s
mast would behave just the same way if the ship were moving steadily
through calm water as it would if the ship were at rest. In both cases the rock
would fall straight down and land at the base of the mast. We express this
idea more generally with the principle of Galilean relativity, which says that
the laws of motion work equally well for anything in uniform motion.
Your immediate surroundings—specifically, those things that participate
in the same motion as you—constitute what physicists call your frame of
reference. Earth is the frame of reference for someone at rest on the planet,
and Galileo’s ship constitutes the reference frame for someone doing
experiments on the ship. Put in terms of reference frames, Galilean relativity
becomes the statement that the laws of motion are valid in any uniformly
moving frame of reference.
Galilean relativity should sound a lot like the idea I introduced in the
preceding chapter: the laws of physics are the same, regardless of one’s state
of motion. That’s essentially what Galilean relativity says, especially
because, in Galileo’s and Newton’s time, many scientists believed that the
laws of motion could explain all of physical reality. So a statement about the
laws of motion becomes, to a Newtonian believer, a statement about all of
physics. Here’s the important point: Galileo and Newton, like you yourself,
already understood and accepted the principle of relativity. For them, as for
you, there was no conceivable experiment that could answer the question Am
I moving? Galileo’s rock dropped from a ship’s mast provides an example:
the rock always lands right at the base of the mast, so the outcome of the
experiment can’t be used to determine whether the ship is moving. Other
questions, such as Is the ship moving relative to the shore?, Am I moving
relative to Earth?, or Is Earth moving relative to the Sun? are answerable—
but these are questions about relative motion. That’s why the statement that
the laws of physics are the same, regardless of one’s state of motion, is a
relativity principle. It implies that motion itself is undetectable and therefore
meaningless; all that matters is relative motion. Ultimately, the relativity
principle traces back to Galileo’s recognition that motion—at least uniform
motion—is a natural state that requires no explanation.
An everyday example should remind you how obvious and simple an
idea is Galilean relativity. Imagine you’re on an airplane, cruising through
calm air at a steady 600 miles per hour. You’re served a snack of airline
peanuts. Do you need to think about the airplane’s 600-mph motion in order
to get the peanuts successfully to your mouth? Do you need to modify your
understanding of the laws of motion to describe the peanuts’ behavior in the
“moving” airplane? Of course not! The airplane’s “motion” is irrelevant.
Newton’s laws work as well on the plane as they do on the ground. That’s
why I put the words “moving” and “motion” in quotes. In the context of
Newtonian physics, it makes no sense to consider that the airplane is
“moving” and Earth isn’t. The laws of physics work the same in both places,
so neither has a claim on some special state of being at rest. If you want to
claim that Earth is really at rest and the plane really moving, then I’ll
challenge you to come up with some physical test that will prove you correct.
Pull down the window shades so you can’t see Earth slipping past, wait for
the plane to be cruising through calm air at constant speed, and then think up
something you can do—that is, a physical experiment—whose results will be
different in the plane than they are if you performed an identical experiment
on Earth. You won’t find it. That’s what it means to say there’s no experiment
that can answer the question, Am I moving? By looking out the window you
can justifiably assert that the plane is moving relative to Earth, but that’s as
far as you’ll get. You’re just as correct in asserting that Earth is moving
relative to the plane. It simply doesn’t matter; with Galilean relativity, there’s
no such thing as absolute motion or absolute rest. This is all just a rehash of
the tennis-match argument from Chapter 2, where the universal scope of the
argument made it even more obvious how absurd it would be to assert that
Earth alone among all the cosmos is truly at rest.
Throughout this book I’ll be making the point again and again that there’s
no experiment you can do to answer the question, Am I moving? or,
equivalently, that identical experiments performed in different reference
frames give exactly the same results. Let me make very clear what this does
and does not mean. What it does mean is that if you and I, each in uniform
motion but each moving relative to the other, set up and perform identical
experiments, then we’ll get exactly the same results. But those experiments
need to be truly identical in all respects except for our relative motion.
That’s why I set the tennis match in Chapter 2 on an indoor court, to
eliminate any wind that would result in an outdoor court from the ship’s
motion relative to the air. That’s why Galileo’s rock-drop experiment on the
ship should really be in an enclosed space, or maybe behind a sail, to block
the apparent wind caused by the ship’s motion. And to be really careful, I
should do the airplane experiment when the plane is at low altitude, to
minimize changes in gravity with distance from Earth. In practice, it may be
hard to set up circumstances that are identical except for relative motion, but
in principle it’s possible. And if we do, then identical experiments will give
identical outcomes.
What my statement does not mean is that observers in relative motion
must see the results of the very same individual experiment in the same way.
If I watch Galileo’s rock-drop experiment from onboard the ship, I’ll see the
rock fall straight down. If you watch the very same experiment from the
shore, you’ll see the rock following a curved path because it shares the
ship’s motion relative to you. But if we both perform different yet identically
staged rock-drop experiments, you on shore and I on the ship, then we will
get identical results—namely the rock falling straight down, taking exactly
the same time as measured by our different but identical clocks, and so forth
down to the last detail. We simply won’t be able to use any of our results to
decide which of us is moving and which isn’t.
We can express the principle of Galilean relativity in many ways. “The
laws of mechanics are the same for all observers in uniform motion” or “The
laws of mechanics are the same in all uniformly moving frames of reference”
are formal ways to state the principle. To say that “I am moving” and “I am at
rest” are meaningless statements is another way of putting Galilean relativity.
“You can eat dinner on an airplane” or “You can play tennis on a cruise ship”
are statements of Galilean relativity applied to specific circumstances.
“Uniform motion doesn’t matter,” “Uniform motion is undetectable,” and
“Absolute motion is a meaningless concept” are still other ways to state the
principle of Galilean relativity.
There are two subtle distinctions between the Galilean relativity
principle of this chapter and my more general introduction of the relativity
principle in Chapter 2. First, Galilean relativity applies only to the laws of
motion, whereas my earlier relativity principle was a statement about all of
physics. To one who believes that motion can explain all of physics, there’s
no difference. But if we should discover new areas of physics that aren’t
based in the laws of motion, then we’ll need to ask anew whether the
relativity principle holds there as well.
The second distinction is that I left out the qualifying phrase “in uniform
motion” in Chapter 2. But we definitely need that qualification in expressing
Galilean relativity. The laws of motion are decidedly not valid if you’re not
in uniform motion. Try eating those peanuts when the plane encounters
unexpected turbulence or while it’s accelerating down the runway. Then the
peanuts act in a decidedly non-Newtonian fashion. They don’t stay put on
your tray table. Try tossing one into your mouth and it goes violently astray,
landing instead on the floor. Or try playing tennis with the cruise ship sailing
through a hurricane. Your intuitive feel for how the ball is supposed to
behave will fail you; more formally, Newton’s laws of motion just don’t hold
in the nonuniformly moving reference frame of the storm-lashed ship. So at
least as far as Galileo and Newton are concerned, statements about
nonuniform motion, that is, changing motion—are meaningful. It does make
sense to say “this airplane’s motion is changing” even though it makes no
sense, in the context of Galilean relativity, to say “this airplane is moving.”
Absolute motion is meaningless in the Galilean/Newtonian view, but change
in motion is supremely meaningful and, indeed, is what Newton’s laws of
motion are all about.
As I move toward Einstein’s special relativity in the next few chapters,
I’ll continue to maintain the distinction between uniform and nonuniform
motion. Later, in Chapter 14, you’ll see how Einstein’s general relativity
blurs that distinction. But the distinction between the laws of motion (with
their associated Galilean relativity principle) and all the laws of physics
will continue to play a major part as we develop first the special and then the
general theory of relativity.
CHAPTER 4

LET THERE BE LIGHT

• • •

Newton’s laws are remarkably sweeping in their scope, explaining

phenomena from the behavior of gases to the game of tennis to the motions of
planets, stars, and galaxies. But are there branches of physics that don’t fall
under the umbrella of Newton’s laws? What about phenomena like sound and
light, our primary ways of communicating with the world around us? Or
electricity, at the heart of modern technology and, more fundamentally,
responsible for the structure and behavior of matter from atoms and
molecules to our own bodies and brains?

Waves

We’ll begin with sound, a classic example of a wave. What’s a wave? Think
of an ocean wave or the wave of standing people that sweeps around a sports
stadium. In each case there’s a disturbance of some medium. For ocean
waves, that medium is the water. For stadium waves, the medium comprises
the people in the stadium. The disturbance moves through the medium,
temporarily upsetting the status quo and then moving on. But the medium
itself doesn’t go anywhere, although it may move briefly back and forth or up
and down as the wave passes by. That is, water from the distant ocean
doesn’t actually move toward shore with the waves. Watch a boat in rough
water: the boat bobs up and down as a wave passes, but it isn’t carried
shoreward with the wave (things get a bit more complicated in the shallow
water where the wave breaks or as a surfer intentionally rides a wave,
moving with the wave by sliding down its sloping front). Similarly, those
sports fans rise from their seats, then return; they don’t go anywhere, even as
the wave circles the stadium.
So what does move in a wave? Certainly the disturbance itself—the
displacement of water from its normal flat surface or the displacement of the
sports fans’ bodies from their sitting positions. In the process the wave
carries energy. In our examples that energy is associated with the temporary
lifting and movement of the water as the wave passes or with the standing
people in the stadium. So a good definition of a wave is that it’s a traveling
disturbance that carries energy but not matter.
Two simple ideas seem inherent in our definition of a wave as a traveling
disturbance—ideas that will play a key role in the development of relativity.
First, since a wave is a disturbance, it seems logical that it must be a
disturbance of something—and that something is what we’ve called the
medium for the wave. Second, since a wave is a traveling disturbance, it
must travel with some speed. What determines that speed? Simple: the
properties of the medium. For water waves, speed depends on things like the
density of water and the force of gravity. In fact, we can deduce that speed by
applying Newton’s laws to the motions of the water, taking account of forces
like gravity and water pressure. So we can add water waves to our growing
list of phenomena explained by Newton’s laws.
Now, on to sound. This, too, is a wave, and one that we can also
understand by using Newton’s laws. A sound wave is a disturbance of the
air, consisting of regions of high and low air pressure (Figure 4.1). The
disturbance moves through the air, but the air itself only moves back and
forth, and very slightly at that. The medium for sound waves is clearly the air.
We can directly measure the speed of sound or we can calculate it by
applying Newton’s laws and accounting for the force associated with the
sound wave’s air pressure variations. Either way, the answer for air under
normal conditions is about 700 miles per hour. That, incidentally, is also
about 1,000 feet per second, or one-fifth of a mile per second, which is why
you count the time between lightning and thunder and divide by five to get the
distance to the lightning strike. Each 5 seconds represents 5,000 feet or about
a mile of travel for the sound waves. Implicit in this rule of thumb is an
assumption that will soon become crucially important to us: the speed of light
from the lightning is so fast that the travel time for the light is negligible.
Fig. 4.1 A sound wave consists of alternating regions of comp ressed (dark) and rarefied (light) air. The entire structure moves
with the sp eed of sound (large arrow). The air itself, however, only moves back and forth as the wave p asses (small arrows).

So we understand sound; it’s a wave ultimately governed by Newton’s

laws applied to the molecules that make up the air. The medium for sound is
air, and the speed of sound is about 700 miles per hour. Seven hundred miles
per hour relative to what? That’s obvious—relative to the air, which is the
medium in which a sound wave is a disturbance. Let me belabor this obvious
point, since we’ll soon be raising the analogous question for light. The speed
of sound is 700 miles per hour relative to the air. If you’re running through
the air at, say, 10 miles per hour, then the speed of sound waves relative to
you should be 710 miles per hour if you’re running toward the source of the
waves and 690 miles per hour if you’re running the other way. No big
mystery here; sound is a disturbance of the air, so to say its speed is 700
miles per hour is to say that sound moves at 700 miles per hour relative to
the air. Sound must move at a different speed relative to someone who is
moving through the air.
I’ll finish with sound by returning to the question that began this chapter:
Are there branches of physics that don’t fall under the umbrella of Newton’s
laws? Sound most definitely is under the Newtonian umbrella, because it’s
entirely explainable by applying Newton’s laws of motion to air molecules.
So sound is solidly a part of Newtonian physics, and we still don’t have an
affirmative answer to our question.

Light: Waves or Particles?

In Opticks, published in 1704, Newton himself gave one of the first thorough
accounts of the nature of light. Inspired perhaps by the success of his laws of
motion, Newton imagined light to consist of particles whose behavior—
including the phenomena of reflection and refraction (the bending of light’s
path as it enters a transparent substance)—could be understood in terms of
the laws of motion. Newton explained color, an obviously important property
of light, by positing that different light particles excite vibrations that give us
the sensation of different colors.
But Newton’s Opticks had competition. In 1678 the Dutch scientist
Christiaan Huygens proposed that light, like sound, consists of waves.
Although Huygens did not address color, a wave theory of light provides a
natural explanation of color in the different frequencies (rates of vibration) of
the waves—just as different frequencies of sound waves correspond to
different pitches. And although the wave and particle explanations were very
different, no experiment at the time could distinguish them. However,
Newton’s stature was such that his particle theory dominated over Huygens’s
waves until the nineteenth century.
There’s one thing waves do that particles cannot. Two waves can pass
through the same place at the same time; when they do, the disturbances that
each represents simply add to make a combined wave. This phenomenon is
called interference. When crests of one wave meet crests of another, the two
waves reinforce to make a larger wave. This is constructive interference.
When crests meet troughs, the interference is destructive and the waves
cancel. Figure 4.2 shows constructive and destructive interference. If the
waves don’t line up perfectly crest-to-crest or crest-to-trough, then the
interference is somewhere between fully constructive and fully destructive.

Fig. 4.2 Wave interference. (a) Constructive interference occurs when crests from two waves meet. (b) Destructive interference
occurs when crests meet troughs.

In 1801 the English physician and physicist Thomas Young performed an

experiment that led ultimately to the downfall of Newton’s particle theory of
light. Young let sunlight pass into a darkened room through two pinholes,
illuminating a screen opposite the holes. If light consisted of particles, one
would expect to find two bright spots on the screen opposite the holes
(Figure 4.3a). Instead, Young saw a series of alternating bright and dark
spots on the screen. It’s difficult to imagine how beams of particles could
produce such a pattern, but the wave theory provides an obvious explanation
(Figure 4.3b). Each hole acts like a source of waves, whose circular crests
spread outward from the holes much like ripples from a rock dropped into a
pond. The waves from the two holes meet and interfere. In some regions,
crests meet crests and troughs meet troughs; here the interference is
constructive and the wave strengthens. In optical terms, that means brighter
light. In other regions, crests and troughs meet. Here the interference is
destructive and the wave is diminished. Figure 4.3b shows that these regions
of constructive and destructive interference lie along lines radiating from the
vicinity of the two holes. Where the lines of constructive interference meet
the screen, bright spots appear. At the lines of destructive interference, the
screen is dark.

Fig. 4.3 Shining light through a p air of small holes in an op aque barrier would p roduce very different effects, dep ending on
whether light consists of p articles or waves. (a) Particles would p roduce two bright sp ots on the screen op p osite the barrier.
(b) Waves p ass through the holes, p roducing circular waves, which subsequently interfere. Thick lines mark regions where
crests meet crests, and troughs meet troughs. Here the interference is constructive, making the light brighter. Bright sp ots
ap p ear where these regions hit the screen. Between, destructive interference results in dark sp ots. To the right is an actual
p hoto of the resulting p attern that ap p ears on the screen.

What determines where the bright and dark spots appear? That depends
on the spacing of the two holes and on the wavelength of the light—the
distance between wave crests. The central bright spot, for example, forms
from waves that have traveled equal distances from the two holes and thus
meet “in step,” with crests joining crests and troughs joining troughs. The
dark zones on either side are from waves that have traveled further from one
of the holes by just the right amount that crests meet troughs, making for
destructive interference. Move the holes closer together and the bright and
dark patches move farther apart, making the interference pattern more
obvious. Move the holes farther apart, and the bright and dark patches get so
close that they soon blur together. The reason the wave nature of light isn’t
obvious to us is that the wavelength of light is very small—around 20
millionths of an inch—so it takes very closely spaced holes to make Young’s
interference pattern noticeable.
So now we have an answer to our question about the nature of light:
Waves or particles? That answer, clear from Young’s interference
experiment, is “waves.” But waves of what? To answer that question takes us
on a new and unexpected turn.

Electricity and Magnetism

The ancient Greeks knew of naturally occurring magnets and their attraction
for iron, and they also studied the attraction of amber for small bits of cloth,
straw, and similar materials—an attraction we now know to be electrical in
nature. The Greeks even speculated on a relation between these two
mysterious forces. But it was not until the late sixteenth century that fruitful
study of electrical and magnetic phenomena began. That study culminated in
the mid nineteenth century with a complete theory of electromagnetism that
then led directly to Einstein’s theory of relativity.
Experimentation with so-called static electricity revealed a new and
fundamental property of matter, electric charge. It was Benjamin Franklin
who clarified that there are two types of charge, which he called positive and
negative. By 1788 the French scientist Charles Augustin Coulomb had
expressed quantitatively the force between electric charges: like charges
repel and opposites attract, with a force that depends on the charges and—as
with Newton’s gravitational force—on the inverse square of the distance
between the charges.
Similar experiments with magnetism established that every magnet has a
north and a south pole and that like poles repel while opposite poles attract.
In contrast to electricity, no one has ever found an isolated north or south
pole, which would be the magnetic analog of electric charge.
 Meanwhile electrical science had advanced with the development of the
battery and the subsequent study of electric current—the flow of electric
charge in wires and electrically conducting solutions like salt water.
Centuries of hunches that electricity and magnetism might be related
culminated dramatically in 1820, with the Dane Hans Christian Oersted’s
discovery that an electric current deflected a nearby magnetic compass. The
French physicists Jean-Baptiste Biot, Félix Savart, and André-Marie Ampère
soon quantified the relation between electric current and magnetism.
Establishing the relation between electric current and magnetism marks
the beginning of one of the great syntheses in human intellectual history—the
joining of all electrical and magnetic phenomena, formerly believed distinct,
under the single umbrella of electromagnetism. The hope for similar
syntheses is a deeply philosophical driving force behind much of physics.
Today, that hope manifests itself in the quest for a “theory of everything” that
would explain all known physical phenomena in a single, coherent scheme.
We now know that electric current or, more fundamentally, moving
electric charge, is the principle source of magnetism. There does not appear
to be any basic, intrinsic property of matter that leads to magnetism in the
way that electric charge leads to electrical phenomena. Rather, magnetism is
simply one aspect of the phenomena associated with the existence of electric
charge. And, again, the phenomena of magnetism arise not from electric
charge alone but from moving charge.
You might be wondering what all this has to do with the familiar magnets
you use to stick notes to your refrigerator. Does their magnetism involve
electric charge? It sure does. The magnetism of everyday magnets results
from the motion of electric charges, in this case the electrons moving about in
the atoms that make up the magnet. Those motions make each atom a
miniature magnet. That, in fact, is true of most atoms. But in a few materials,
notably iron, the individual atomic magnets interact in such a way that their
magnetic effects reinforce, resulting in the large-scale magnetism we see in
ordinary permanent magnets. There’s really not much difference between a
permanent magnet and an electromagnet—one made by winding a coil of
wire and passing electric current through it. In both cases the magnetism
results from the motion of electric charges.
The relation between moving charge and magnetism is a two-way street.
Moving charge not only creates magnetism but it also responds to magnetism.
Fundamentally, the interaction of magnetism with matter results from a force
that magnets exert on moving charge. Place an electron—a fundamental
constituent of matter that carries negative charge—near a magnet and nothing
happens. But if it’s moving, the electron experiences a force that depends on
the strength of the magnet and on the electron’s motion. That’s why your TV
has circuitry to compensate for Earth’s magnetism. The same terrestrial
magnetism that guides compass needles would alter the paths of the electrons
that beam through your TV tube to create the picture you see on the screen.
Once again, magnetism is about moving electric charge.
Are there other relations between electricity and magnetism? In the 1820s
the English physicist Michael Faraday pursued this question and by 1831 he
had an answer. Because electric current produced magnetism, Faraday
wondered if, conversely, magnetism might produce electric current. He made
an electromagnet by winding a coil of wire and passing electric current
through it. He placed another coil nearby and connected it to a meter that
would register a flow of current in this second coil. Nothing happened. But
then Faraday noticed a deflection of the meter when he interrupted the current
in the first coil. That interruption represented a change in the current and
hence in the magnetism of the first coil. Further experimentation convinced
Faraday that he had discovered a new phenomenon, wherein changing
magnetism can produce electric current. Today we call this phenomenon
electromagnetic induction, and we exploit it in huge generators that spin
wire coils in the presence of magnets to generate the electricity that powers
our homes and industries. We also use it to read information from our
videotapes and computer disks, whose changing magnetic patterns induce
currents in a wire coil placed near the moving tape or spinning disk.
Implicit in the preceding paragraphs are four fundamental statements
about electricity and magnetism. I’ll state them very loosely here, but later
we’ll take a more careful and complete look at them:

• Electric charges exert forces on each other.

• Magnetic poles exert forces on each other, but there are no isolated
magnetic poles.
• Moving electric charge gives rise to magnetism.
• Changing magnetism gives rise to electricity.
These four statements look rather different, but they actually comprise two
pairs that reflect a profound symmetry between electric and magnetic
phenomena. We need to develop that symmetry more fully in order to answer
our question about the nature of light.

Fields

Suppose Earth suddenly vanished. How and when would the Moon, a
quarter-million miles away, “know” that it should abandon its circular orbit
—the result of Earth’s gravity—and begin the straight-line motion that
Newton’s laws tell us it should take in the absence of a force? In Newton’s
view, the Moon would know about Earth’s disappearance instantaneously,
because in Newton’s description of gravity Earth “reaches out” across empty
space and immediately “pulls” on the Moon. Remove Earth, and that pull—
the gravitational force—disappears instantly. This Newtonian idea is called
action at a distance for obvious reasons.
There’s another way to view the Earth–Moon interaction, a way that at
first may seem needlessly complicating and abstract. But it will be crucial in
developing your understanding of light and will become absolutely essential
in the context of relativity. Even at this early stage, you may find this new
view more philosophically satisfying.
Here’s the issue: The action-at-a-distance description of gravity requires
the Moon to know what’s happening, right now, at the location of the distant
Earth. But how can this be? It would be much more believable if the Moon
only needed to know about its immediate vicinity and responded only to
local conditions. Enter the concept of field. We imagine that Earth creates a
kind of influence, called a gravitational field, everywhere in the space
around it. Put an object in Earth’s vicinity and it experiences not a mysterious
action-at-a-distance pull from Earth but rather the gravitational field right at
the point where the object is. It responds to that field by experiencing a force
toward Earth’s center, a force whose strength depends on the object’s mass
and on the strength of the gravitational field. Since gravity decreases with
distance from Earth, so must the strength of the field. We can represent the
gravitational field by drawing arrows that show its strength and direction at
selected points (Figure 4.4).
 Conceptually, in introducing the field concept we’ve replaced the direct
but philosophically disturbing action-at-a-distance force of gravity with a
force that arises locally, from the gravitational field at any point in space.
Rather than Earth exerting forces directly, we have the more complex
situation in which Earth creates a gravitational field in its vicinity and
objects respond to that field. What we gain from this complexity is a new
simplicity: now an object doesn’t have to know about the situation at some
distant place but only about what’s happening in its immediate vicinity.
Of course, the ultimate outcome remains unchanged. The Moon, a
spacecraft, and a falling apple still behave as Newton predicted. Only our
description of how that occurs has changed; now these objects respond to
Earth’s gravitational field right where they are, rather than to Earth itself. But
for now the action-at-a-distance and field perspectives predict exactly the
same physical results.
Now back to the question What if Earth suddenly vanished? We can’t
answer that question for sure at this point, but at least the field concept gives
us some wiggle room. What happens at Earth’s location isn’t important to the
Moon; it’s the local gravitational field that determines the Moon’s motion. So
does the gravitational field everywhere vanish instantaneously when Earth
does? Or does it have some sort of independent existence so it takes awhile
for the field way out at the Moon’s location to learn about Earth’s demise? In
fact, the answer is the latter, as we’ll explore in later chapters. For now,
though, suffice it to say that the field concept gives us a new and important
physical entity—in this case the gravitational field—that can have, at least
temporarily, an independent existence.
Fig. 4.4 A rep resentation of Earth’s gravitational field. The length of each arrow gives the strength of the gravitational field at
its location, and the arrows p oint in the direction of the field, namely, toward Earth’s center. The field exists every where, but I
can only show arrows to rep resent it at a few selected p oints.

Electromagnetic Fields

I could have phrased the entire preceding section in terms of the electric
force between charges rather than the gravitational force between Earth and
Moon. I chose the latter only because you’re more familiar with gravity.
Having established the concept of the gravitational field, though, the electric
field follows directly by analogy. We say that one electric charge creates an
electric field everywhere in space and that a second charge responds to the
field in its immediate vicinity. Again we get the same physical happenings as
predicted by the action-at-a-distance electric force, but now the behavior of
a charge is determined by local conditions—in this case the local electric
field. Again we can speculate about what would happen if one charge
suddenly vanished or just moved a bit; would a second, distant charge
respond immediately or would it take a while for the local field to learn
about this change and thus to adjust to it?
If electric charges create electric fields, surely magnets must create
magnetic fields. And they do. Given our understanding of magnetism, though,
we can put all this more carefully: Electric charges create electric fields,
while moving electric charges create, in addition, magnetic fields. And we
can rephrase Faraday’s discovery as a statement about fields: the loose
phrase “changing magnetism gives rise to electricity” becomes “A changing
magnetic field creates an electric field.” This last statement has an interesting
implication: no longer is electric charge the sole source of electric
phenomena, in particular of electric fields. Now something quite different,
namely a changing magnetic field, can also create an electric field. So
electric fields have two possible sources: electric charge and changing
magnetic fields. Notice that the second of these entails a direct relation
between the fields themselves, where one directly creates the other without
the intermediary of electric charge or any other matter.
Let’s look in more detail at the last two of our earlier statements about
electromagnetism, now recast in terms of electric and magnetic fields:

• Moving electric charge creates a magnetic field.

• A changing magnetic field creates an electric field.

There’s something vaguely unsatisfying, something unsymmetrical, about

these two statements. If a changing magnetic field creates an electric field,
might not a changing electric field create a magnetic field? That question
occurred to the great Scottish physicist James Clerk Maxwell in the mid
nineteenth century. Maxwell had good reason to suspect that a changing
electric field would indeed create a magnetic field. First, our statements
when expressed as mathematical equations look much more similar than the
English versions, and the one virtually cries out for a mathematical term
meaning “changing electric field” where the other has a term “changing
magnetic field.” Second, Maxwell saw that the equations were not fully
consistent with the known fact that electric charge is conserved. He found
that he could regain consistency by adding a term “changing electric field” in
the statement/equation about the source of magnetic field. So he did, positing
that his modified set of four statements provides a complete and consistent
description of all electromagnetic phenomena. In Maxwell’s honor, we now
call those four statements collectively Maxwell’s equations.
Electromagnetic Waves

So what’s the big deal? Now a changing electric field creates a magnetic
field and, as we knew from Faraday’s work, a changing magnetic field
creates an electric field. The big deal, as Maxwell soon deduced, is that this
ability of each field to create the other results in a wondrous new
phenomenon. Suppose, somehow, a changing electric field gets started. We
might make such a field by grabbing an electron or other electrically charged
object and moving it back and forth. That changing electric field creates a
magnetic field. In general, that magnetic field will also be changing, so it
creates an electric field. In general, that electric field will also be changing,
so it creates a magnetic field. And so forth. You get the picture: once a
changing field of either type appears, a self-perpetuating system of electric
and magnetic fields is set into existence. Once they get started, the fields take
on an independent existence, no longer associated with the electric charge or
whatever it was that got them started.
In fact, Maxwell realized, this self-perpetuating structure of electric and
magnetic fields wouldn’t just sit there; rather, it would propagate through
space as an electromagnetic wave. Like the other waves we’ve examined,
such a wave would carry energy but not matter. Now Maxwell was not
particularly interested in practical applications for these electromagnetic
waves; he just wanted a complete, consistent description of electromagnetic
phenomena. But in fact he had discovered something immensely useful. In
1887 the German Heinrich Hertz, again in the spirit of pure scientific inquiry,
sought to confirm Maxwell’s theory. He set up an electric circuit with a
power source that made electrons wiggle back and forth, creating a changing
electric field. A few feet away sat another circuit with no source of power.
When the first circuit was energized, a spark jumped in the second circuit—
indicating the arrival of energy via Maxwell’s electromagnetic waves. Thus,
Hertz vindicated Maxwell’s theoretical hunches.
Was this work practical? Not yet. But in the 1890s the Italian-Irish
physicist Guglielmo Marconi took up the challenge and used Maxwell’s
electromagnetic waves to establish so-called wireless telegraphy as a
practical means of communication over long distances. In 1901 a triumphant
Marconi sent electromagnetic signals from England to North America. Our
modern radio, television, cell phones, satellite dishes, microwave ovens,
electronic car keys, wireless computer mice, police radar, garage door
openers, radio telescopes, and countless other technologies are all based on
electromagnetic waves. All owe their existence to the wondrous interactions
between electric and magnetic fields as described by Maxwell’s equations.
My brief history of electromagnetic wave technology was just an aside to
help you appreciate the immense practical significance of Maxwell’s work.
An even more philosophically profound result awaits us next.

Let There Be Light

Maxwell not only recognized the possibility of electromagnetic waves, but

he also explored their properties as predicted by his equations. Among those
properties is the speed of electromagnetic waves, which Maxwell could
easily calculate from his equations. Those equations contain references to the
electric and magnetic fields, of course, and also to electric charge and
electric current. But they also contain two numbers that set the overall
strength of the electric and magnetic forces. These two numbers represent
fundamental properties of nature. Their values were determined
experimentally, the electrical number from experiments on the electric force
between charged objects and the magnetic number from experiments with
electric circuits and the magnetism they produce. Those experiments, done in
the eighteenth and nineteenth centuries using the simple apparatus of the time,
were all that was needed to establish the values of the electric and magnetic
numbers that appear in Maxwell’s equations.
When Maxwell proceeded to calculate the speed of his electromagnetic
waves, he found that it depended only on the values of these electric and
magnetic numbers—quantities that, again, had been determined in simple
laboratory experiments involving electric charge and electric current. When
he worked out the value for the speed of his waves, Maxwell found it to be
very nearly 300 million meters per second, or about 186,000 miles per
second—and equal to the speed of light! That speed had been known since
the late seventeenth century, and by Maxwell’s time it was solidly
established to an accuracy of a few percent. So there was no question: the
speed of Maxwell’s electromagnetic waves was the speed of light.
Maxwell’s conclusion was inescapable: light itself is an electromagnetic
wave! In one brilliant stroke, Maxwell brought the whole science of optics
under the umbrella of electromagnetism.
What a remarkable achievement is Maxwell’s! First, he establishes a
fully consistent theory of electromagnetism, expressed succinctly in the four
Maxwell equations. He finds that his theory predicts an entirely new
phenomenon—electromagnetic waves. Next he calculates the speed of those
waves and recognizes that it’s equal to the known speed of light. So now,
some 60 years after Young’s interference experiment established that light
consists of waves, Maxwell has the answer to our question, Waves of what?
Light consists of waves of electricity and magnetism, or electromagnetic
waves.
How is light related to the electromagnetic waves of Hertz and Marconi
and of modern communications? Simple: They’re essentially the same thing.
The only difference is the frequency with which the wave vibrates, or
equivalently, the wavelength (the distance between wave crests, which
decreases as the frequency increases). I’ve already mentioned that light
waves have a wavelength about 20 millionths of an inch. AM radio, in
contrast, has a wavelength of about 300 yards, while the waves used in FM
radio have a wavelength of about 3 yards. The microwaves that cook your
food have a wavelength of 3 inches, and police radar operates at a
wavelength of about one-fifth of an inch. In fact, there’s a whole continuous
spectrum of electromagnetic waves, starting with the longest wavelength
radio waves on down through the microwaves used for cooking, radar, and
satellite communications. Shorter still and we’re in the infrared—
electromagnetic waves with wavelength slightly longer than those of visible
light. You can feel infrared waves from a hot stove burner, clothes iron,
woodstove, and other hot objects. Special infrared cameras can image “hot
spots” indicating disease in the body or energy loss in a house. Still shorter
and we’re in the visible range of electromagnetic waves, from about 25
millionths of an inch for red light to about 14 millionths of an inch for deep
violet. Beyond that is ultraviolet, which you can’t see but which produces
sunburn; then x-rays, and finally the penetrating gamma rays produced in
nuclear reactions. We give all these ranges of electromagnetic waves
different names out of convenience because of the different uses they have
and the different means of producing them, but fundamentally they’re all the
same: all are electromagnetic waves, consisting of self-perpetuating
structures of electric and magnetic fields, and all propagate at exactly the
same speed. That speed, which we call the speed of light and denote by the
symbol c, is really a universal speed of all electromagnetic waves. (That c
comes most likely from the Latin celeritas, meaning swiftness.)
One caveat: c is the speed of electromagnetic waves in a vacuum. When
electromagnetic waves travel through matter, they interact with electric
charges in the matter (because light is an electromagnetic wave) and this
slows down the wave. The phenomenon of refraction, for example, occurs
because light moves more slowly in the plastic of your contact lenses or the
glass of your camera lens than it does in air or vacuum. When I use the term
“speed of light” in this book, I’ll always mean the speed of light in vacuum
unless I explicitly state otherwise. The speed of light is such a fundamental
quantity that it’s useful to know its numerical value. Again, that value is very
nearly 300 million meters per second; 186,000 miles per second (more than
seven times around the Earth in 1 second); or, in units appropriate to high-
speed computers, about 1 foot per nanosecond (billionth of a second). It’s
also, in units useful to astronomers and in discussing relativity, exactly 1
light-year per year. More on this later.

Making Electromagnetic Waves

How do we make electromagnetic waves? How does nature do it? What we
need to get a wave started is a changing field, either electric or magnetic. So
in principle we could make an electromagnetic wave by taking an
electrically charged object or a magnet and moving it back and forth, or
spinning it, or whatever. The frequency of the wave would be the frequency
of our back-and-forth or spinning motion and that would also determine its
wavelength. (A detailed look at Maxwell’s equations shows that the charge
or magnet must be accelerating, that is, it must be in a state of changing
motion. Uniform motion won’t do, for reasons that relativity will make
obvious.)
In practice, we make the relatively low-frequency electromagnetic waves
in the radio region of the electromagnetic spectrum using electric circuits that
feature rapidly alternating electric currents flowing in the antennas of radio
transmitting systems. In your microwave oven and in radar units, electrons
circle around billions of times a second in a special device called a
magnetron, in the process emitting electromagnetic waves because circular
motion is accelerated motion. The frequency of these waves is that of the
electron’s circular motion and corresponds to wavelengths of several inches
and shorter. The still higher frequency, shorter wavelength infrared waves
arise naturally from vibrating molecules. And most visible light, as well as
ultraviolet, comes from individual atoms, as electrons jump about among
different energy levels in the atom. In a medical x-ray unit, high-speed
electrons slam into a metal target; their abrupt stopping constitutes the
acceleration (change in motion) that produces electromagnetic waves in the
x-ray region of the spectrum. Finally, energetic and extremely rapid
processes within atomic nuclei produce the so-called gamma rays,
electromagnetic waves with the highest known frequencies and shortest
wavelengths. Again, the names we give these different electromagnetic
waves are just for convenience; all the waves are basically the same, being
self-perpetuating structures of electric and magnetic fields that propagate at
the speed of light, c. Only the frequencies and corresponding wavelengths
distinguish the different types of waves.

A Brief History of Physics

So here we are, in the late 1800s, with what looks like a complete
understanding of all aspects of physical reality. That understanding springs
from two different branches of physics: the mechanics (laws of motion) of
Newton and the electromagnetism of Maxwell. (Of course many others were
involved in establishing both branches, but for brevity I’ll use only the names
of Newton and Maxwell.) Many phenomena, from the behavior of gases to
the motion of planets to the properties and propagation of sound, have
Newtonian explanations. Electromagnetic phenomena, including all of optics
(the study of light), follow from Maxwell’s equations. Historically, Newton’s
mechanics came first, with the years from roughly 1600 to 1750 seeing the
greatest advances in the Newtonian understanding of the Universe.
Electromagnetism followed, with major advances from 1750 to nearly 1900.
Together, these two distinct branches seemed to explain all of physics.
 But do they really explain all of physics? And are their explanations
consistent? Recall that Newtonian mechanics obeys a relativity principle—
meaning that Newton’s laws work equally well in any uniformly moving
frame of reference. Is the same true of Maxwell’s laws of electromagnetism?
You now have enough background that these and related questions can lead
you straight to an understanding of Einstein’s thought.
CHAPTER 5

ETHER DREAMS
• • •

You’ve just seen how Maxwell’s electromagnetism leads to electromagnetic

waves, including light, and predicts that these waves should move with
speed c, about 186,000 miles per second. But 186,000 miles per second
relative to what? In the preceding chapter we asked the analogous question
for sound, Seven hundred miles per hour relative to what?, and had an
obvious answer. Sound moves at 700 miles per hour relative to air, the
medium in which sound waves are a disturbance. So what’s the
corresponding answer for light? (When I say “light” in this context, I mean
any electromagnetic wave, including visible light. I’ll often use “light” and
“electromagnetic wave” interchangeably.)
Before we try to answer this question, let me emphasize why it’s so
important. Without an answer, Maxwell’s equations are on shaky ground
because their prediction of electromagnetic waves moving with speed c is
meaningless. If we can’t say with respect to what that speed c is measured,
then the statement “Light goes at speed c” is pretty vacuous. We’re now going
to explore possible answers to the question, Speed c relative to what? If you
find you don’t like those answers, then I challenge you to come up with your
own! Keep in mind that we need an answer if Maxwell’s electromagnetic
theory—a theory that is supposed to explain a great many important physical
phenomena including much of modern technology—is to have a solid
grounding.
Enter the Ether

Nineteenth-century physicists, heady with the successes of Newtonian

mechanics, naturally assumed that electromagnetic waves are like sound and
all other known waves, namely, that they are disturbances of some medium.
They called that medium the ether. Then, Speed c relative to what? had an
obvious answer: Light goes at speed c relative to the ether. An immediate
corollary is that if you’re moving through the ether, then light’s speed relative
to you won’t be c—just as the speed of sound waves measured by someone
moving through the air is different from the 700-miles-per-hour-speed of
sound relative to air.
Let’s take a closer look at this ether. Ether is supposed to be to
electromagnetic waves what air is to sound waves—the medium in which the
waves are a disturbance. Now sound waves are possible only where there’s
air, which is why astronauts on the airless moon can’t communicate by
talking. But light reaches us from the far corners of the Universe, so ether
must be everywhere. Ether must be a kind of tenuous, transparent
“background” that pervades the entire Universe. It fills the space between the
planets, stars, and galaxies. Because light propagates through transparent
substances like air and water, ether must permeate the tiny spaces between
and even within atoms.
This ubiquitous ether would need some unusual properties. It would have
to be a fluid like air or water, rather than a solid, since things move through
it. Further, it should be a very tenuous fluid, because it offers no resistance to
the motion of planets. If it did, Earth and the other planets would lose energy
and eventually spiral into the Sun. But at the same time ether would need to
be very stiff, because the speed of light is so big. You can see this if you
imagine a stretched spring, Slinky, or rubber band. Send a wave pulse down
the spring by disturbing it briefly and then letting go. The more you stretch the
spring, the faster the pulse travels. An analogous situation holds for sound
waves in different media; sound travels faster in water than in air, for
instance, because water is “stiffer” in the sense of the spring with more
stretch. And an analogy holds for electromagnetic waves in the ether. For
those waves to have the high speed of 186,000 miles per second, the ether
must be very stiff indeed. The two ether properties of being tenuous, to
minimize resistance, and being stiff, to maximize wave speed, aren’t easily
reconciled. Ether would have to be a rather unusual material.

Don’t Like the Ether? Try Another Answer

Its contradictory properties make ether an improbable substance, but so
ingrained was the mechanical paradigm and so crucial is it to answer the
question, Speed c relative to what? that nineteenth-century physicists saw no
alternative to the ether. You may think you see an alternative, namely, do
away with the ether entirely and let electromagnetic waves propagate through
empty space. If you go down that road, then I’m going to ask you for an
answer to the question Speed c relative to what?, and I’m going to expect a
good answer that’s consistent with physical experiments.
Here’s a seemingly good answer: Light goes at speed c relative to its
source. Eliminate the ether, that is, and let light propagate through empty
space at 186,000 miles per second relative to whatever object emitted the
light. An observer at rest with respect to that light source would measure c
for the light’s speed, but an observer for whom the source was moving would
obviously measure a speed different from c. This idea is pretty simple. An
analogy would be a baseball pitcher who can throw a 100-mile-per-hour
fastball. Thrown from the pitcher’s mound, the ball moves at 100 miles per
hour relative to the pitcher. Since the pitcher is standing on the ground, that
amounts to 100 miles per hour relative to Earth. But put the pitcher in an
open car going 50 miles per hour. Again, the ball goes 100 miles per hour
relative to its source, namely the pitcher. But relative to observers on the
ground, it would be going 150 miles per hour. (Or so it seems; more on this
later!)
So does the measured speed of light depend on the speed of its source
relative to the observer? No, it doesn’t. A host of experiments and
observations confirm this. A particularly straightforward confirmation comes
from astronomical observations of double-star systems. Unlike our Sun,
about half the stars in the Milky Way galaxy are locked in a gravitational
embrace with another star. These double stars orbit around each other with
orbital periods ranging from hours to years. In some cases the system lies
with the plane of its stars’ orbits in line with our view from Earth. Then as
the stars orbit, one moves toward us while the other moves away; half a
period later this situation is reversed (Figure 5.1). We can analyze the stars’
motions in detail by studying slight shifts in the wavelength (i.e., color) of
light as we observe it. These changes arise because of the star’s motions
relative to Earth.
Now even the nearest stars are very far away—so far that light from them
takes many years to reach us. A slight difference in the speed of light from the
two stars in a double-star system would show up as a long time delay
between our receiving light that was, in fact, emitted simultaneously from the
two stars. We would have to compensate for this difference in light travel
times to infer the stars’ motions. But we need no such compensation! The
stars move exactly as Newton’s law of gravity says they should, based on the
light as we receive it. There’s no increase in the speed of light from the star
that’s momentarily moving toward us and no decrease for the star that’s
moving away from us. The speed of light simply does not depend on the
motion of its source.
So what’s wrong with our baseball analogy? What’s wrong is that light
doesn’t consist of particles like miniature baseballs. Rather, as you know, it
consists of waves. Think about sound waves for a minute. The speed of
sound waves is 700 miles per hour relative to the medium, namely air, in
which sound is a disturbance. If a source of sound—say a fire truck’s siren—
moves through the air at 60 miles per hour, the speed of the sound is still 700
miles per hour relative to the air. It doesn’t pick up the additional 60 miles
per hour of the fire truck. Apparently light is like that. It must move at c
relative to the medium, namely ether, in which it is a disturbance. If a source
of light moved through the ether, the speed of the light would nevertheless
still be c relative to the ether.
Fig. 5.1 A double-star sy stem, showing the two stars orbiting each other. At the moment p ictured, the left-hand star is
ap p roaching Earth and the right-hand star is receding. If the sp eed of light dep ended on the sp eed of its source, light from the
left-hand star would reach Earth sooner.

(By the way, something about the sound from a moving source does
change, namely, its pitch. When the fire truck is coming toward you, the siren
sounds like it has a higher pitch. When it moves away, the pitch you hear is
lower. The analogous shift for light is what lets us measure motions in
double-star systems.)
In disabusing you of the notion that the speed of light might be c relative
to its source alone, I reverted naturally to the language of ether—the medium
in which light waves are purportedly a disturbance. Despite the problematic
aspects of the ether, what else are we (or, more precisely, nineteenth-century
physicists) to do? Light, after all, consists of waves, and all other known
waves are disturbances of some medium. The speed of each type of wave—
sound waves, water waves, stadium waves, earthquake waves, etc.—is its
speed relative to its particular medium. Why not the same for light?
 If you’re still troubled by the all-pervasive, tenuous, light-wave-
supporting ether, then again I challenge you to tell me relative to what light
travels at speed c. That’s a harder challenge now you know that the answer
cannot be “relative to its source.” So for now we’ll continue to follow the
nineteenth-century physicists’ line of thought, picturing light waves as
disturbances of an ether that pervades the entire Universe. If you really don’t
like the ether, though, hold on—eventually you’ll be vindicated!

A Broader Question

We were led to the ether concept by questioning relative to what light travels
at speed c. I want to convince you now that this question is really a special
case of a more general one: In what frame of reference are the laws of
electromagnetism (i.e., Maxwell’s equations) valid? The two questions are
related because one prediction of the laws of electromagnetism is that there
should be electromagnetic waves and that they should go at the speed of light,
c. When we answer “ether” to the question “Relative to what does light go at
speed c?”, we’re saying that Maxwell’s prediction of electromagnetic waves
that go at c is really valid only in a frame of reference at rest with respect to
the ether. Observers in a reference frame moving through the ether will
measure some other speed for light relative to themselves; thus for them, in
the context of their reference frame, the predictions of Maxwell’s equations
won’t be valid. That’s why my two questions, one about the speed of light
and the other about the validity of Maxwell’s equations, are essentially
equivalent. Maxwell’s equations predict electromagnetic waves going at c,
so those equations can only be valid in a frame of reference where one will,
in fact, measure c for the speed of light.
So what’s the answer to our new question, In what frame of reference are
Maxwell’s equations valid? It’s obvious: In the context of nineteenth-century
physics, there’s only one frame of reference in which Maxwell’s prediction
about electromagnetic waves is valid, and that’s a frame of reference at rest
with respect to the ether.

Dichotomy in Physics
In Chapter 3 we discovered the principle of Galilean relativity, which states
that Newton’s laws of motion are valid in all uniformly moving frames of
reference. So if we ask explicitly in what frame of reference Newton’s laws
of motion are valid, then the answer is “in any uniformly moving frame of
reference.”
This question we’ve just asked about Newton’s laws of motion is exactly
the same question we asked in the previous section about Maxwell’s laws of
electromagnetism. But there we found quite a different answer. Maxwell’s
electromagnetism, it seems, isn’t valid in just any frame of reference. Rather,
the laws of electromagnetism should be valid only in one very special frame
of reference—a frame of reference at rest with respect to the ether. Put
another way, the laws of motion obey a relativity principle but the laws of
electromagnetism seem not to. So although there’s no experiment we can do
with the laws of motion to answer the question, Am I moving?, there should
be electromagnetic experiments that can answer this question. That is, the
concept of absolute motion is meaningless for mechanics, but apparently it
has meaning for electromagnetism. Although there’s no privileged state of
motion for mechanics, there seems to be a privileged state for
electromagnetism—namely, being at rest relative to the ether.
Why the dichotomy? Why should one main branch of physics (mechanics)
not care about states of motion, while the other (electromagnetism) does?
Wouldn’t it be simpler and more coherent if both branches of physics obeyed
the relativity principle, or both didn’t?
Think back to the second chapter, where I asked about a tennis match
played on a cruise ship, on Venus, and on a planet in a distant galaxy moving
away from Earth at nearly the speed of light. You wisely and logically agreed
that it made perfect sense to expect that tennis playing—a manifestation of the
laws of motion—would work the same in all those contexts or, as we would
now say, in all those different reference frames. That is, you intuitively
accepted the principle of relativity as applied to the laws of motion. Then I
asked about heating a cup of tea in a microwave oven, and you agreed that
the microwave oven, like the tennis ball, should also behave the same way in
the different reference frames. But it’s electromagnetism, not mechanics, that
governs the microwave oven. Through our exploration of the question,
Relative to what does light go at speed c?, we’ve just found that the laws of
electromagnetism seem not to obey the relativity principle. That is, the laws
of electromagnetism should not be valid in all reference frames—and
electromagnetic experiments should therefore give different results in
reference frames that are in different states of motion. So the microwave
oven shouldn’t work the same way in that distant galaxy moving at nearly c
as it does on Earth!
What’s wrong with your intuition from Chapter 2? Again, the difficulty is
with the ether. When you blithely agreed that the microwave oven should
work the same way everywhere, you weren’t taking the nineteenth-century
view, with its unique reference frame of the ether, in which frame alone the
laws governing the oven should be valid.
So why not get rid of the ether, now troublesome not only because of its
improbable characteristics but because it’s also the cause of an illogical and
unsatisfying dichotomy between two branches of physics? That dichotomy
runs counter to your good sense that the laws governing physical reality,
whether tennis balls or microwave ovens, should work the same everywhere
and regardless of one’s state of motion. If you abandon the ether, then you
could eliminate that dichotomy.
But if you abandon the ether, then I challenge you once again to answer
the question, Relative to what does light go at speed c? or, equivalently, In
what frame of reference are the laws of electromagnetism valid? If you give
the same answer that I’ll happily accept for the laws of motion—“in all
uniformly moving reference frames”—then you’ll vindicate your intuitive
sense from Chapter 2 that the microwave oven should work the same in all
states of motion. But if you give that answer, you’ll find yourself on the edge
of a philosophical abyss. That’s because you’ll be insisting on a seeming
contradiction: that two different observers must each find valid the
Maxwellian prediction that light waves move at speed c—even if those
observers are moving relative to each other! Better not go there, at least not
yet; instead, we’ll stick for now with the nineteenth-century ether concept and
explore further its implications.

Earth and Ether

Following nineteenth-century physicists, we’ve established ether as the
medium through which light waves propagate with speed c, and we recognize
therefore that Maxwell’s equations of electromagnetism can only be valid in
a frame of reference at rest with respect to the ether. So a logical question
arises: How is our Earth moving relative to the ether?
It’s pretty obvious that we aren’t moving very fast through the ether,
because if we were then we would see obvious differences in the speed of
light coming from different directions. For example, if Earth were moving at
90 percent of c, then light from stars in the direction toward which Earth is
moving would be going at 1.9c relative to us (c + 0.9c = 1.9c). But light from
stars in the opposite direction would be going at only one-tenth of c (c – 0.9c
= 0.1c). That difference would be patently obvious. But if Earth were moving
much more slowly through the ether—at a tiny fraction of c—then we might
not notice the difference unless we looked very carefully. So again the
question: How is Earth moving relative to the ether? That is, how fast and in
what direction is Earth’s motion relative to the ether?
We begin with an even simpler question: Is Earth moving relative to the
ether? That one has a yes-or-no answer. Either Earth is moving relative to the
ether, or it isn’t.

Insulting Copernicus

Consider first the possibility that Earth isn’t moving relative to the ether. I
can think of only two ways for this to be the case. First, the ether might be a
fixed substance that extends throughout the Universe. Then Earth alone among
all the cosmos would be at rest relative to the ether. I say “alone” because all
other celestial objects—the Moon, Mars, Venus, the other planets, the Sun,
other stars in our galaxy, and the other galaxies in the Universe—all are
moving relative to Earth. So if Earth is at rest relative to the ether, then it
alone is at rest. That makes us pretty special. If we’re the only beings at rest
relative to the ether, then Maxwell’s equations are valid only for us, and only
we measure c for the speed of light. Observers on other celestial bodies
measure different speeds for light in different directions, and for observers
moving very fast relative to Earth—like those in distant galaxies—that effect
must be dramatically obvious.
Copernicus would turn in his grave! It’s hard to imagine a worse insult to
the Copernican revolution than to make our planet so special that one of the
two main branches of physics is valid only on Earth. I spent most of Chapter
3 presenting a history of science that led steadily away from the notion of
Earth being a special, privileged place in the Universe. Do you really want
to return to parochial, pre-Copernican ideas? Do you really think you and
your planet are so special that, in all the rich vastness of the Universe, you
alone can claim to be “at rest”?
On purely philosophical grounds, we should reject the notion that Earth
alone could be at rest relative to the ether. Now, philosophy isn’t science,
and I hasten to add that there’s plenty of good scientific evidence to support
this view. For example, we observe light-emitting processes in distant stars
and galaxies that seem to work the same there as they do here on Earth. That
suggests we don’t have any special status vis-à-vis the laws of
electromagnetism. So we can confidently reject the idea that Earth alone is at
rest relative to the ether.

Ether Drag

It might still be possible for Earth to be at rest relative to the ether if our
planet somehow “dragged” the surrounding ether with it. Presumably other
planets and celestial bodies would do the same, so each would be at rest
relative to its local blob of ether. Then observers everywhere and in different
states of motion would find the laws of electromagnetism to be valid, and no
one would have any claim to be special. Copernicus would be a lot happier
with that!
So does Earth drag the ether with it? Astronomical observations dating to
1725 provide a clear answer. A simple analogy will help you understand
these observations. In Figure 5.2a you’re standing, holding an umbrella in the
rain. Obviously the best approach to keeping dry is to hold the umbrella
directly overhead. But what if you run through the rain? Now it’s better to tilt
the umbrella, holding it at an angle (Figure 5.2b). Figure 5.2c shows why:
viewed from your frame of reference, the rain is coming down at an angle,
and you want to hold the umbrella so the rain still hits the umbrella top
straight on. Since the rain is falling at an angle, you should hold the umbrella
at the same angle. If you run in the opposite direction, you should still hold
the umbrella tilted at the same angle in front of you, but now this will be a
different absolute direction.
However, suppose that somehow you drag a big blob of air with you as
you run—so big a blob that rain falling into it has time for the force of the
moving air to accelerate it to your running speed before it hits you or your
umbrella. Figure 5.2d shows the situation from your point of view. As you
run through the rain, the rain outside your blob of dragged air falls at an angle
as seen from your reference frame. But inside the blob, the moving air
accelerates the rain until it shares the blob’s motion. So now, relative to you,
it’s falling straight. The upshot is that you don’t have to tilt your umbrella.
Rather, you’ll stay driest if you hold it right overhead.

Fig. 5.2 An analogy for the aberration of starlight. (a) Standing still in vertically falling rain, y ou hold y our umbrella straight
overhead to keep driest. (b) Running, y ou tilt y our umbrella. (c) The situation in (b), shown from the runner’s frame of
reference. In this frame, the rain falls at an angle. (d) If the runner drags a large blob of air, then rain entering the blob will take
on the blob’s motion, and thus will fall vertically relative to the runner.

You can see from the foregoing discussion that an umbrella in a rainstorm
is a useful device for determining whether you drag a big blob of air when
you run—as if you don’t already know the answer. But let’s suppose you
don’t. So you set out in a rainstorm to answer the question. You run through
the rain and hold your umbrella first over your head, then at an angle. You
find out which approach keeps you driest. If the overhead approach is best,
then, as in Figure 5.2d, you can conclude that you do drag a big blob of air
with you. But if tilting the umbrella works best, then from Figure 5.2c it’s
obvious that you don’t drag the air with you.
Running in the rain with an umbrella provides an analogy for
astronomical observations that answer the question, Does Earth drag a blob
of ether with it? The runner moving through the air is like Earth moving
through the ether. The rain is like light from a distant star “falling” on Earth.
The umbrella is an astronomer’s telescope. And the question is, Do we need
to point the telescope at an angle to compensate for Earth’s motion? Actually,
the question is slightly more subtle, since we can’t stop Earth and compare
the telescope angles with Earth stopped or moving. But what we can do is
compare the angles for different directions of Earth’s motion. Relative to the
Sun, Earth is right now moving in some direction. Six months later, halfway
around its orbit, the planet will be going in the opposite direction. If Earth
doesn’t drag the ether, then we’ll need to change the angle of our telescope to
observe the same star at six-month intervals, just as running in different
directions in Figure 5.2c requires changing the direction of the umbrella’s
shaft. Such a change in the apparent direction to a star is known as
aberration of starlight. If, on the other hand, Earth does drag the ether, then
we keep the telescope in the same direction for both observations. In this
case there’s no aberration. So what happens? In fact, the telescope angle
changes—a small change, to be sure, but readily detectable with
astronomical measurements as early as 1725. We conclude that Earth does
not drag ether with it.
It seems we’ve answered the question of whether Earth is moving
relative to the ether. On philosophical as well as observational grounds we
first ruled out the possibility that Earth alone is at rest relative to the ether.
Now aberration of starlight shows that Earth doesn’t drag a blob of ether
with it. Those were the only ways we could conceive of Earth’s being at rest
relative to the ether. So we have to conclude that Earth is moving relative to
the ether.
Given that fact, we can now ask how fast and in what direction Earth is
moving. That’s the question nineteenth-century physicists set out to answer.
With a variety of clever experiments, they sought to measure Earth’s motion
through the ether. These experiments typically involved the propagation of
light through water and had the added benefit of verifying the notion that
moving substances—for example, flowing water—do not communicate their
motion to the surrounding ether. That is, the experiments confirmed the
absence of ether drag. But, curiously, they failed to detect Earth’s motion
through the ether. This disappointing result was attributed to the effects of the
water on light, effects that apparently just cancelled out the sought-after
indication of motion through the ether. The question of Earth’s motion
remained ambiguous.

Wrap-up: Physics at 1880

It’s now about the year 1880, and here’s how things stand. Together,
Newton’s mechanics and Maxwell’s electromagnetism seem to explain all
known physical phenomena. There’s a philosophically disturbing dichotomy,
in that a relativity principle holds in mechanics but not in electromagnetism,
but that doesn’t diminish the explanatory power of these two great branches
of physics. Light is understood as an electromagnetic wave, propagating with
speed c through a Universe-permeating medium called ether. Astronomical
observations show that the speed of light does not depend on the motion of its
source and that Earth must be in motion relative to the ether. The only thing
remaining to solidify the picture of electromagnetic waves in the ether is to
measure Earth’s motion. But as of 1880 no experiment has succeeded in
doing so. Then again, the experiments are difficult, they require great
sensitivity, and there are complicating effects that may obscure the desired
result. What’s needed is a conceptually simple experiment that’s sensitive
enough to measure unambiguously Earth’s motion through the ether. Such an
experiment should, once and for all, lay to rest any nagging doubts about the
ether.
CHAPTER 6

CRISIS IN PHYSICS
• • •

How are we to determine Earth’s motion relative to the ether? If light moves
with speed c through the ether then, as I’ve stressed before, observers who
are moving relative to the ether should measure values other than c for the
speed of light relative to themselves. So the most obvious way to detect
Earth’s motion through the ether is to measure the speed of light and see if it
agrees with the value c predicted from Maxwell’s equations. If it doesn’t,
then the observer must be moving relative to the ether, and the difference
between the measured speed of light and c should reveal the speed of that
motion. Unfortunately, no measure of light’s speed in the nineteenth century
was remotely precise enough to distinguish between the actual value c and a
slightly different value caused by a modest earthly motion through the ether.
However, a moving observer will find not only that the speed of light is
not quite c but also that light’s speed depends on the direction of the
observer’s motion in relation to that of the light. As Figure 6.1 shows, an
observer moving toward a source of light will measure a higher speed than c,
while an observer moving away will measure a lower speed. So a more
subtle approach to the problem is to try to detect differences in the measured
speed of light in different directions. This approach avoids having to know
the actual value with extreme precision; all that’s important is measuring
directly the difference in two values, without measuring the values
themselves.
Fig. 6.1 Observers moving relative to the ether (gray rectangles) would measure different sp eeds for light. Vertical lines are
crests of a light wave, and the length of the arrow rep resents the light’s sp eed relative to the observer. (a) A stationary observer
would measure c, the normal sp eed of light. (b) An observer moving toward the source of the light would measure a higher
sp eed, and (c) an observer moving away, a lower sp eed.

Can we get any idea of how much difference to expect? Not entirely,
because we don’t know Earth’s speed relative to the ether. That, after all, is
what we’re trying to find out. But we know that Earth orbits the Sun at some
20 miles per second, and because the orbit is circular, the direction of that
motion changes throughout the year. So even if the Sun, by some bizarre
coincidence, happened to be at rest in the ether, Earth would be moving at
some 20 miles per second. And whatever Earth’s speed, physicists knew it
had to change by some tens of miles per second throughout the year, as Earth
first heads in one direction at 20 miles per second relative to the Sun and
then, 6 months later, at 20 miles per second in the opposite direction. If they
could build a device to detect a speed of that magnitude, they could answer
the question of Earth’s motion through the ether.
Now, 20 miles per second sounds fast but it’s slow compared with the
speed of light, some 186,000 miles per second. Worse, it turns out that a
successful measurement of Earth’s motion requires detecting the square of
Earth’s speed in relation to the square of the 186,000-mile-per-second speed
of light. That’s like measuring the difference between the numbers 1 and
1.00000001. So great was the challenge that many nineteenth-century
experimenters thought it impossible.

Michelson, Master of Light

Enter the Prussian-born American physicist Albert A. Michelson.
Michelson’s passion was the speed of light, and by 1879, while a young
instructor at the U.S. Naval Academy, he had measured that important
quantity to within 0.05 percent of its exact value. His expertise at precision
measurement gave Michelson hope that he might be the first to measure
unambiguously Earth’s motion relative to the ether. Michelson transformed
that hope into series of experiments, culminating in an 1887 version that
remains among the most famous experiments in all of science. I’m going to
spend some time discussing this experiment because it provides some of the
most convincing evidence in support of Einstein’s relativity. (For you, and
for most physicists, that is. Whether the experiment influenced Einstein
himself remains a matter of debate among historians of science. More on this
later.)
Michelson was a brilliant and meticulous experimenter, and he invented a
device that, to this day, forms the heart of instruments providing precise
measurements of distance, time, and other quantities. This device, the
Michelson interferometer, uses the interference of light waves to detect
minute differences in the time light takes to travel two different paths.
Michelson’s interferometer can easily measure time differences less than the
oscillation period of light—about one-thousandth of a trillionth of a second.
With that kind of precision, Michelson knew he could detect variations in the
speed of light due to Earth’s orbital motion.
Conceptually, Michelson’s experiment is simple. Two beams of light
travel equal-length paths oriented at right angles to each other. Because of
Earth’s motion through the ether, the apparatus experiences an “ether wind,”
just as you feel a wind when you stick your hand out the window of a moving
car, even on a calm day. Since light travels at speed c through the ether, this
ether wind will affect the time it takes light to travel through Michelson’s
apparatus. Because the two beams travel paths that are oriented differently,
the wind’s effect on the two beams will, in general, be different. The idea is
to detect and measure that difference.
I’ll consider in detail only the simplest case, in which one light path lies
along the direction of the ether wind (i.e., along the direction of Earth’s
motion through the ether) and the other lies at right angles. Now, for reasons
that will soon be obvious, Michelson’s light beams travel on round-trip
paths. So the beam moving along the direction of the ether wind has to move
both upstream, against the wind, and then downstream, with the wind. Going
upstream, the light is slowed as measured by an observer on Earth, just as a
boat traveling up a flowing river moves more slowly relative to the bank than
if the river weren’t flowing. Going downstream, with the ether wind, the
light’s speed will increase. So will the boat’s. Now, you might think these
two effects cancel out—but they don’t. Here’s why: Because it’s moving
slowly going upstream, the light (or the boat) spends more time being slowed
by the ether wind (or the river current) than it does being sped up on the
return trip. In fact, if the ether wind were blowing at the speed of light c, the
light would never get anywhere and its round-trip would take forever!
Similarly, if you rowed a boat upstream at just the speed of the current, you
wouldn’t get anywhere in relation to the riverbank. So the light’s
upstream/downstream trip will always take longer than it would in the
absence of an ether wind.
Meanwhile, the other light beam is taking a round-trip journey at right
angles to the wind. You might think the wind would have no effect on this
beam, but that’s not so. Again imagine a boat, which you’re now trying to
row to a point straight across a river. If you aim straight across, you won’t
get where you want to be, because the current will drag you downstream. So
you have to aim a bit upstream and that makes the trip take a little longer. You
have to do the same on your return trip, so in this case, too, the current slows
the boat. But because on this perpendicular trip the current isn’t affecting you
as directly, the slowing-down effect is less than in the upstream/downstream
trip. (A little work with ninth-grade algebra and the Pythagorean theorem
could convince you rigorously of this conclusion.) So in Michelson’s
apparatus the light whose path lies along the ether wind should take longer to
make its round-trip than the light whose path is perpendicular to the wind.
That time difference is what Michelson set out to measure.
Even with Earth’s 20-mile-per-second orbital speed, the time to travel
Michelson’s two paths is tiny—measured in billionths of a second. And the
expected time difference would be about a million times smaller, so much so
that even if one could measure the times, one couldn’t possibly do so with
enough accuracy to determine their difference. Here’s where Michelson’s
genius comes in. Instead of trying to measure the two times, he went directly
for the time difference—which, after all, is the significant manifestation of
the ether wind. It was the nature of light itself—the wave nature of light—that
let Michelson measure that time difference. Because light consists of waves,
it exhibits interference—the phenomenon that, back in 1801, had convinced
Thomas Young of light’s wave nature. In describing Young’s experiment in
Chapter 4, I distinguished between two types of interference, constructive
and destructive. These led, respectively, to the bright and dark bands in the
pattern of light falling on Young’s screen. Now here’s the crucial point:
constructive interference (bright light; wave crests in step) turns into
destructive interference (dark; crests meeting troughs) if one of the two
interfering light waves gets delayed just enough that its troughs fall back to
where its crests had previously been. Because light’s wavelength (distance
between crests) is so small, it doesn’t require much of a time delay to turn
constructive interference to destructive. How much? About a thousandth of a
trillionth of a second!
So here’s Michelson’s idea, as shown in Figure 6.2. Produce a single
beam of light of a pure color (meaning a single wavelength). Split the beam
in two, and send the resulting beams on two equal paths at right angles.
How? With a half-silvered mirror, also called a beam splitter. This is just a
lousy mirror that didn’t get enough reflective coating. So only about half the
light that hits it gets reflected, while the other half goes straight through.
Michelson placed his beam splitter at a 45-degree angle to the original light
beam. The result, shown in Figure 6.2, is to send half the light straight on and
to reflect half of it perpendicular to its initial direction. Both light beams then
travel on equal-length paths, bounce off ordinary mirrors, and return to the
half-silvered mirror. Again, half of each beam is reflected and half goes
through. The important point, as Figure 6.2 shows, is that portions of each
beam come together again, traveling toward the viewer at the bottom of the
picture.
Fig. 6.2 In the M ichelson–M orley exp eriment, a light beam sp lits and travels two p erp endicular p aths. Light returning from the
mirrors is rejoined at the beam sp litter, resulting in interference seen in the viewer. Photo below the viewer is the interference
p attern from an actual M ichelson–M orley setup .

If the beam paths in Figure 6.2 were exactly equal, and if there were no
ether wind, then the beams would return exactly in step and the light waves
would interfere constructively. An observer looking into the viewer would
see bright light. But suppose the ether wind delayed one beam by exactly
enough to make its troughs line up with the other beam’s crests. Then we
would have destructive interference and darkness in the viewer. That, in
principle, is how Michelson’s interferometer could detect Earth’s motion.
Actually, it’s both a bit more complex but also simpler than that. First of
all, it’s impossible to get the two ordinary mirrors at exactly right angles, and
in any event the light rays in the beams aren’t exactly parallel but diverge
slightly and hit the half-silvered mirror at slightly different points. The result
is that light takes a whole lot of different paths, differing slightly in length, on
each of the two legs of Michelson’s apparatus. Some of it returns in step,
interfering constructively, and some of it returns out of step. The result is not
simply bright or dark in the viewer but a pattern of alternating light and dark
bands. (Figure 6.2 includes a photo of interference bands from a modern-day
Michelson interferometer). Furthermore, it’s impossible to get the distances
to the two mirrors exactly the same. But all that does is to alter just which
light interferes constructively and which destructively. Even if the path
lengths aren’t the same, we’ll still get a pattern of light and dark bands. It’s
just that the positions of the bands will be a bit different.
None of these subtleties matters, though; in fact, they make the experiment
easier, since we don’t have to worry about getting the path lengths equal or
the mirrors exactly perpendicular. The reason they don’t matter is that we
aren’t really interested in the interference pattern itself, which includes the
effects not only of the ether wind but also of the various imperfections in the
instrument like path lengths and mirror alignment. But now suppose we rotate
the whole apparatus through 90 degrees. The path that was initially along the
ether wind is now at right angles to it, and the one that was originally at right
angles is now along the wind. Whatever the relative timing for the light
beams on the two paths was originally, the important point is that the timing
should now change. That change is due entirely to the ether wind because
nothing about the apparatus itself has changed except its orientation in the
wind. And how do we detect that change in the relative travel times on the
two paths? Simple: we watch the interference pattern. A change in travel
times should result in a shift in the positions of the bright and dark
interference bands. That shift, viewed as the apparatus is rotated, gives a
direct indication of Earth’s motion through the ether. The magnitude of the
shift is a measure Earth’s speed.

The Michelson–Morley Experiment

By 1880 Michelson, then on a 2-year study trip to Europe, had built his first
interferometer. With this device, he knew, Earth’s orbital motion should
produce a shift of only a few percent of the distance between bands—barely
detectable. Returning to the United States, Michelson secured a position at
Case School of Applied Science in Cleveland. He soon began collaboration
with Edward Morley, a chemist from nearby Western Reserve University
(much later merged with Case to form Case Western Reserve University). By
1887—after enduring Michelson’s brief but forced commitment to a “nerve
specialist” for supposed “softening of the brain” and a disastrous fire at the
Case School—Michelson and Morley had ready a much improved version of
Michelson’s interferometer. This device was mounted on a 5-foot square
stone slab floating in liquid mercury, allowing easy rotation of the entire
apparatus without disturbing the delicate light paths. Use of multiple mirrors
greatly increased the effective lengths of the two light paths. As a result the
Michelson–Morley experiment of 1887 was sensitive enough that a shift of
nearly half the distance between light and dark fringes should occur as a
result of Earth’s orbital motion—a shift that would be obvious to the
observer, as shown in Figure 6.3.

Fig. 6.3 M ichelson and M orley should have seen a shift of nearly half a fringe, p utting a dark band where a light band had been.
Photos, from an actual M ichelson–M orley setup , are offset to show the fringe shift.

Michelson and Morley performed their experiment by slowly rotating the

apparatus as one of them walked around with his eye to the viewer. They
repeated the experiment at different times so their entire lab would be
oriented differently relative to the ether wind. And what was the result? A
dismal failure: they never saw any significant shift in the interference bands!

Contradiction!

The starkly negative outcome of the Michelson–Morley experiment stands as

one of the most important experimental results in all of science. To see why,
remember where we are, logically, in 1887. The realization that light is a
wave, specifically an electromagnetic wave propagating at speed c, raises
the question of the medium in which light propagates. Nineteenth-century
physics answers that question by proposing the ether as the medium for light
and other electromagnetic waves. It then makes sense to ask about Earth’s
motion relative to the ether. Aberration of starlight shows that Earth can’t be
at rest relative to the ether, so Earth must be moving. Earlier experiments fail
to detect that motion, but they suffer either from conceptual flaws or
insufficient sensitivity. Now, in 1887, come Michelson and Morley with an
experiment much more sensitive than what’s needed to detect something that
must exist, namely, Earth’s motion through the ether. And yet the experiment
fails. Earth must be moving through the ether, yet the Michelson–Morley
experiment shows that it isn’t. That’s a pretty stark contradiction, and it
shook the foundations of physics in the late nineteenth century.
I repeat: it’s the foundations of physics—the basis of our whole
understanding of physical reality—that are shaking, not some minor
inconsequential details. Why is this contradiction so profound, so dire?
Because it concerns a fundamental and sweeping prediction of one of the two
basic branches of physics, specifically Maxwell’s electromagnetism with its
prediction of electromagnetic waves propagating at speed c. That prediction
immediately gives voice to the question, Speed c relative to what? It’s in
attempting to answer that question that the contradiction inherent in
Michelson–Morley arises. If we can’t resolve that contradiction, then there’s
something drastically wrong with our supposed understanding of physical
reality.
No wonder nineteenth-century physicists sought at all costs to explain
away the negative Michelson–Morley result. Not to do so would be to admit
a logical fault so deeply ingrained as to threaten the entire edifice of physics.
The experiment itself seemed beyond reproach, so physicists sought some
explanation, some excuse, for its failure to detect Earth’s motion. Michelson
himself concluded disappointedly that Earth must be at rest relative to the
ether after all, despite the apparently opposite implication of the starlight
aberration observations. Others made more radical suggestions. In particular,
the Dutch physicist H. A. Lorentz and the Irish physicist George Fitzgerald
independently proposed that the ether squeezes objects moving through it,
contracting them in the direction parallel to their motion. This contraction
would shorten the ether-wind-aligned path in the Michelson–Morley
experiment and thus reduce the travel time for light along that path. If the
contraction were just right, the effect would eliminate the time differences on
the two paths and would therefore explain the negative result of the
experiment.
The Lorentz–Fitzgerald contraction would be very small, amounting to a
decrease of only about 3 inches in Earth’s 8,000-mile diameter. But that’s all
it would take to explain the Michelson–Morley result and to restore the
logical consistency of physics in the nineteenth century. But given ether’s
tenuous nature, why should the contraction occur at all? And why should it be
the same for all substances, regardless of what they’re made of? There was
no satisfying answer to these questions, and the Lorentz–Fitzgerald
contraction seemed a very ad hoc way out of the Michelson–Morley
contradiction.
Ingenious though the nineteenth-century physicists were, they weren’t
quite ingenious enough to break free of their nineteenth-century mindset and
resolve the contradiction in a simple, fresh, and radical way. That resolution
had to wait until 5 years into the next century, and when it came it was truly
revolutionary yet profoundly simple.
CHAPTER 7

EINSTEIN TO THE RESCUE

• • •

Albert Einstein was 8 years old when Michelson and Morley performed their
1887 experiment. Slow to speak, socially withdrawn, and stormy of temper,
the young Einstein seemed not especially promising. But a magnetic compass
his father showed him had evoked in 5-year-old Albert the first glimmers of
his fascination with the deepest nature of physical reality. The compass
needle’s mysterious response to something unseen and unfelt made a lasting
impression and stirred Einstein’s lifelong search for the hidden principles
governing the physical Universe.

Einstein: Approaching the Magic Year

Einstein the student got mixed reviews. In elementary school he was strong in
logical subjects like mathematics and Latin, but showed little motivation for
others that failed to captivate him. As an older student, Einstein’s interests
broadened to include philosophy, literature, and music. Always, though, he
rebelled against rote learning, and eventually dropped out of his
militaristically disciplined German high school to join his family who were
then living in Italy. Without completing high school, and 2 years younger than
the minimum age for admission, he took the entrance exam for the Zurich
Polytechnic Institute. He did impressively well in mathematics and physics,
but failed other subjects. Nevertheless, he was admitted with a year’s delay,
contingent on his gaining a high-school diploma. So Einstein enrolled in a
high school in Aarau, Switzerland.
 During his year at Aarau, at age 16, Einstein had a prescient insight that
shows his mind was already stretching toward relativity. He wondered what
would happen if one ran alongside a light wave, at its same speed.
Obviously, one should see a stationary structure of electric and magnetic
fields—just as a sailor moving with the speed of water waves would see,
relative to the boat, stationary crests and troughs of water. The trouble was,
Einstein realized, that a stationary electromagnetic wave structure is not
consistent with Maxwell’s electromagnetic theory. Much later Einstein
remarked that 10 years’ speculation on this quandary helped lead him to
relativity.
In 1896, Einstein enrolled at the Zurich Polytechnic Institute. Here he
continued his iconoclastic approach to education. Disgusted with a professor
whose lectures were so outdated that they didn’t include Maxwell’s
electromagnetism, Einstein skipped class and studied Maxwellian theory on
his own. He was arrogant and confrontational to some of the physics faculty,
who in turn found him insolent and lazy; they even suggested that he switch to
medicine or literature. Einstein also found time for love with his fellow
physics student Mileva Mari´c . She had come to Zurich to study medicine,
then switched to physics—an unlikely choice for a woman in her day, but
made possible by the liberal climate at Zurich, which had earlier graduated
the first female PhDs in Europe. Mari´c ’s academic record was comparable
to Einstein’s, and in 1900 they and three classmates all took the final exam.
Their grades were the lowest of the five; Einstein passed, but Mari´c didn’t.
Nevertheless, Einstein’s letters reveal, they looked forward to a life of
collaborative work in physics.
Success did not come easily. After graduation from Zurich, Einstein
failed to find a permanent position. His classmates had all been offered
faculty appointments at Zurich, but Einstein had so antagonized his professors
that they would have none of him. Then Mileva discovered she was pregnant.
Their daughter Lieserl was born early in 1902, but soon all trace of her
vanished from the historical record. Did she die young? Did she live,
perhaps never knowing herself to be Einstein’s daughter? There’s simply no
evidence. Einstein, meanwhile, tried unsuccessfully to support his family as a
private tutor. Eventually he secured a technical job in the Swiss Patent
Office. This lowly position was actually good for Einstein; in addition to a
stable income and contact with interesting technological ideas, it gave him
plenty of time to pursue his own work in physics.
Einstein and Mileva Mari´c married in 1903. In 1904 their son Hans
Albert was born. That brings us to 1905.

The Magic Year

Asked to picture Einstein, you probably conjure up images of an old man
with wild hair and a frumpy sweater. But that’s not the Einstein of relativity.
In 1905 Einstein was a young father of 26 years, devoted to his family, to his
work at the patent office, and to his physics (Figure 7.1). Despite his busy
young life, lack of an academic position, and a paucity of colleagues,
Einstein in 1905 nevertheless produced a burst of scientific advancement
unprecedented except, perhaps, for Newton’s plague years at Woolsthorpe.
Fig. 7.1 In 1905 Einstein was a y oung father, shown here with his infant son Hans Albert. (Courtesy of The Albert Einstein
Archives, The Hebrew University of Jerusalem, Israel.)

In 1905 Einstein completed a total of six scientific papers. Three of these

stand out as seminal works in the history of physics. The first paper
introduced the photon, or quantum of light energy, and helped establish the
fledgling science of quantum physics. That work earned Einstein the 1921
Nobel Prize in Physics. Two papers dealt with the size of molecules and
provided the final convincing evidence that matter really does consist of
atoms and molecules. The first of the molecular papers was Einstein’s PhD
dissertation for the Zurich Polytechnic; ironically, he submitted it only after
the Zurich faculty had rejected his paper on relativity! The fifth paper of the
year—a mere footnote to the theory of relativity—introduced the idea behind
the famous equation E = mc2.
 Then there’s the fourth paper. And a remarkable one it was. It cited no
other scientific papers, contained a minimum of mathematics, and made few
references to specific experiments. Yet with incisive clarity, ingeniously
penetrating insight, and utter simplicity, Einstein in this paper resolved
completely the contradictions posed by Michelson–Morley and other
attempts to answer the question of Earth’s motion.
Was this paper “The Theory of Relativity”? Yes, in content. No, in title.
Rather, its modest title is “On the Electrodynamics of Moving Bodies.” That
title emphasizes the intimate connection between electromagnetism and
relativity. Relativity grew out of contradictions arising from Maxwell’s
revelation that light consists of electromagnetic waves propagating at speed
c. The natural question, Speed c relative to what? followed directly. The
nineteenth-century answer, ether, then led to questions about Earth’s motion.
Recall that it also led to a perplexing dichotomy, in which one branch of
physics—Newton’s mechanics—obeys the relativity principle, meaning that
the laws of mechanics are valid in all uniformly moving reference frames.
But the other branch—Maxwell’s electromagnetism—seemed to be valid
only in the ether’s reference frame and therefore did not obey the relativity
principle.
Einstein’s resolution was at once conservative, radical, and profoundly
simple. It’s stated in one brief sentence, called the Principle of Relativity:

The laws of physics are the same in all uniformly moving reference
frames.

That’s it. This one sentence implies all of Einstein’s special theory of
relativity. It’s conservative because it asserts for electromagnetism—indeed,
for all of physics—what had been known for centuries about mechanics,
namely, that there’s no favored state of motion, no preferred frame of
reference. If you look back in Chapter 3, where I introduced the principle of
Galilean relativity, you’ll find that Einstein’s statement is identical except
that the phrase “laws of motion” becomes generalized to “laws of physics.”
So in this sense Einstein’s relativity is nothing new. It’s just a generalization
of Galilean relativity to all of physics—including electromagnetism.
Einstein’s relativity is clearly simple, stated in its entirety in just one brief
sentence, but it’s also radical. For Einstein to state it, and for you to
understand it fully, requires dramatic restructuring of deep-seated
commonsense notions about the nature of space and time. That is what the
rest of this book is about, and it’s what makes relativity seem a challenging,
even mind-boggling subject. But I want to stress again and again that, in
essence, relativity is simplicity itself. It’s all based on the fact that there’s no
preferred state of motion for describing physical reality; that you and I, as
long as we’re both moving uniformly, will discover exactly the same
underlying physical laws—even though we may be moving relative to each
other. That’s exactly the idea that I introduced in Chapter 2, when I convinced
you that a tennis match and a microwave oven should work the same on
Earth, on Venus, and on that planet in a distant galaxy moving away from
Earth at nearly the speed of light. Motion—as long as it’s uniform motion—
simply doesn’t matter. There’s no one who can claim, “You’re moving and
I’m not.” All states of motion—at least uniform motion—are equally valid.
There’s no such thing as being absolutely at rest or in motion. Only
statements about relative motion make sense.
By the way, you might be questioning that caveat about uniform motion.
Why shouldn’t all states of motion be equivalent? We’ll get there eventually,
as Einstein did by 1915. But first things first: Einstein’s 1905 relativity is the
special theory of relativity. Special here doesn’t mean that the theory is
particularly great and wonderful (although it certainly is both!), but special in
the sense of being specific and limited. Special relativity is limited to the
special case of reference frames in uniform motion. Einstein’s general
relativity, which I’ll introduce in Chapter 14, removes that limitation.
(Historically, Einstein actually presented relativity in the form of two
postulates. The first was his assertion that the Principle of Relativity, which I
stated above, applies to all of physics. The second, which Einstein noted “is
only apparently irreconcilable” with the Principle of Relativity, asserts that
the speed of light is the same in all uniformly moving reference frames. A
more modern approach takes the view that the second postulate follows from
the first; indeed, by 1910 physicists had shown rigorously that the second
postulate is superfluous.)
Here’s how the Principle of Relativity takes care of those pesky
nineteenth-century dilemmas and contradictions. The Principle asserts that all
the laws of physics are the same in all uniformly moving reference frames.
Among the laws of physics are Maxwell’s equations of electromagnetism.
Those equations lead to the prediction that there should exist electromagnetic
waves and that those waves should propagate at a particular speed—the
speed of light, c. Speed c relative to what? Nineteenth-century physicists
foundered around in the ether trying to answer this question, without success.
But the Principle of Relativity provides a simple answer. Because the
prediction of electromagnetic waves propagating at speed c is a prediction of
the laws of physics, and because those laws are valid in all uniformly
moving reference frames, it must be the case that electromagnetic waves
propagate at speed c as measured in any uniformly moving reference frame.

Farewell to the Ether

Einstein’s answer to the question, Speed c relative to what? renders the

concept of ether unnecessary and, indeed, dangerously misleading. The ether
was supposed to provide a medium in which light waves could propagate,
and as such it would constitute a frame of reference that could claim to be
truly at rest and in which alone the laws of electromagnetism would be valid.
But the idea of a preferred frame runs afoul of the Principle of Relativity. If
the laws of physics are valid in all uniformly moving reference frames, then
the idea of a preferred ether frame has to go. So there is no ether. Relativity
does away with it once and for all. As Einstein himself put it in his 1905
paper, “the introduction of . . . ether will prove to be superfluous.” Light
propagates not through some material medium but through empty space. In
that sense light waves are different from all other waves physicists had
encountered, all of which require a physical medium—air for sound, water
for water waves, rock for earthquake waves, and so forth.

Have Faith in Relativity!

There’s a disturbing implication of relativity’s assertion regarding the speed

of light. Observers in different reference frames must still get the same value
for the speed of light, even though they’re moving relative to each other.
That means you and I, measuring the speed of the very same light, get the
same answer even though I’m at rest on Earth and you’re whizzing by in a
car, an airplane, or a high-speed rocket. It’s that consequence—the
invariance of the speed of light, even for observers who are moving relative
to one another—that’s so troubling and that’s going to lead to a radical
revision of your commonsense notions of space and time. We’ll explore these
ideas thoroughly in the next chapter.
For now, though, concentrate on where the invariance of c comes from. It
results from nothing more than the Principle of Relativity, as applied to the
laws of electromagnetism with their prediction of electromagnetic waves
(including light) going at speed c. That’s all. There’s nothing hidden. Accept
the Principle of Relativity—as you did intuitively with the tennis and
microwave examples of Chapter 2—and the invariance of the speed of light
follows. When you find yourself in disbelief at some consequence of
relativity in subsequent chapters, just come back to this simple point. If you
accept the Principle of Relativity, then it’s just two steps to the invariance of
c: (1) The laws of electromagnetism predict electromagnetic waves going at
speed c and (2) The laws of physics are valid in all reference frames; thus,
the conclusion that electromagnetic waves go at c must be valid in all
reference frames. From the invariance of c follow the many seemingly
counterintuitive results of special relativity.
But why should you accept the Principle of Relativity? As a reader with
a twenty-first-century perspective on the vastness of our expanding Universe,
you readily bought my argument in Chapter 2 that the tennis match and the
microwave oven should work the same on Earth, on Venus, and on that
distant planet moving rapidly away from Earth. If you want to question that,
ask yourself if you really believe Earth to be so special that only here, or,
more accurately, only in a reference frame at rest with respect to Earth, are
the laws of physics valid. In all the vast Universe, with countless galaxies
each containing billions of stars and planets whizzing apart in the cosmic
expansion, could our state of motion be so special?
Even if you want to ignore your post-Copernican instinct that Earth and
its state of motion aren’t special, look at the experimental evidence and the
struggle of nineteenth-century science to fathom the question of Earth’s
motion. As I detailed in Chapters 5 and 6, the nineteenth century closed with
no satisfactory answer to this question that is consistent with observations
like aberration of starlight and experiments like Michelson–Morley. But
Einstein’s 1905 answer is consistent and simple: The laws of physics are the
same in all uniformly moving reference frames. Period. Although conclusions
drawn from Einstein’s simple statement are going to upset your notions of
space and time, I’ll emphasize now and again later that those conclusions,
while strange, are never contradictory. They follow from only one simple
fact: The laws of physics are the same in any uniformly moving reference
frame.
CHAPTER 8

STRETCHING TIME
• • •

Why is relativity so hard to swallow? Because the invariance of the speed of

light—a straightforward consequence of the Principle of Relativity—leads
directly to what seem impossible conclusions. In this chapter I’ll use the
invariance of c in reaching these strange conclusions, but don’t come away
thinking that relativity is just about the behavior of light. Relativity is
ultimately about the absence of any favored reference frame in the physical
Universe, and all the strange conclusions we’ll draw follow from that idea.
The invariance of the speed of light is only one of many consequences of the
relativity principle. The behavior of light provides a convenient stepping
stone to relativity’s deeper implications, but keep in mind that those
implications reflect the fundamental nature of space and time, and aren’t just
about light.

Measuring the Speed of Light

Figure 8.1 shows you standing by a road. A traffic signal some distance away
emits a flash of light, and the figure shows the light waves as they come by
you. How fast? At speed c, of course. You could, in principle, verify that
speed if I gave you a meter stick and a very accurate stopwatch. You would
start the watch when the light flash first passed the front end of your meter
stick and stop it when it passed the back end. Divide the distance—1 meter
—by the measured time, and you’ll have the speed of light. With a perfect
meter stick, a perfect stopwatch, and perfect experimental technique, you’ll
get precisely the value c = 299 792 458 meters per second (this is the exact
value for c; my “186,000 miles per second” and “300 million meters per
second” are approximate). Of course you can’t do this experiment with
ordinary meter sticks and stopwatches, but modern electronics provides
timing devices sufficiently fast and light pulses sufficiently short that one can
easily measure c precisely over distances of a few meters—the size of a
typical room.

Fig. 8.1 Four observers each measure the same value c for the sp eed of light relative to themselves, even though they ’re in
motion relative to one another.

Figure 8.1 also shows a friend driving by in a car at 60 miles per hour,
toward the traffic signal. I’ve given your friend a perfect meter stick and a
perfect stopwatch, identical to yours. Your friend uses exactly the same
experimental technique on exactly the same light flash, to determine the speed
of the flash. And what does she find? Because she’s heading toward the
traffic signal, you might guess that she’ll measure a slightly higher speed than
your 299 792 458 meters per second—higher by her speed of 60 miles per
hour, or about 27 meters per second. But that answer isn’t consistent with the
Principle of Relativity, which states that the laws of physics are the same in
all uniformly moving reference frames. One consequence of those laws is the
existence of electromagnetic waves—including light—that propagate at
speed c. Both you and your friend are in uniform motion, so that particular
consequence of the laws of physics must hold for both of you. Both of you
must measure the same value for the speed of light c. With perfect meter
sticks and stopwatches, that value will be 299 792 458 meters per second.
 Let me make it perfectly clear that I’m saying that each of you gets
exactly the same value for c. You might think your speed relative to your
friend is so small that you just don’t notice the difference. So to make things
more dramatic, Figure 8.1 shows two more observers, one in a jet plane
going at 600 miles per hour and another in a rocket moving toward the traffic
signal at half the speed of light. Surely an astronaut on the rocket sees the
light approaching at 1.5c. But no! That result would be inconsistent with
relativity for the same reason it would be for the driver of the car. You, your
friend, the airplane pilot, and the astronaut are all in uniform motion and,
therefore, the laws of physics are equally valid for all of you. None of you—
including yourself standing on the road—can claim in any absolute sense to
be at rest. None of you can say that the laws of physics are correct only for
you. You’re all in equally good situations—reference frames—for exploring
physical reality, and one consequence is that you’ll all measure precisely the
same value for the speed of light, c.
But how can this be? How can different observers, moving relative to
one another, still measure the same speed for the same light? Ultimately the
answer is simple: The laws of physics are the same for all observers in
uniform motion, and the invariance of c follows directly from that principle.
Whenever you find yourself asking, How can this be? about relativity, the
answer always lies in the Principle of Relativity. Remind yourself how
intuitive that principle seemed in Chapter 2, and how the experimental
evidence and scientific quandaries of the nineteenth century led Einstein to
affirm the relativity principle as the basis of his theory. If you still don’t like
what relativity has to say, then you answer the question, Speed c relative to
what? in a way that’s consistent with experiments like Michelson–Morley
and that doesn’t put Earth, alone among all the cosmos, in a favored position.

A [False] Analogy with Sound

You might be tempted to use an analogy with sound to argue against my
conclusion that all four observers in Figure 8.1 measure the same speed for
light, so let’s explore that analogy. Suppose we replace the traffic signal with
a loudspeaker, a musical instrument, or a person shouting. Out comes a burst
of sound waves, heading for the four observers. They measure the speed of
this sound, each with the same meter stick and stopwatch technique (this is
easier because sound goes much more slowly than light). Doesn’t each
observer get a different value for the speed of sound? Surely your friend in
the car, heading toward the source of sound, measures a higher value than you
do—higher by her speed of 60 miles per hour. Because a commercial jet’s
600 miles per hour is nearly the speed of sound, the airplane pilot should
measure almost twice the speed you do. And to the astronaut—woosh!—the
sound should seem to be going by at about half the speed of light. So is that
what happens? Do the different observers measure different speeds for the
sound waves? If they do, doesn’t that invalidate the conclusion that they’ll all
measure the same speed for light waves?
Yes, but no. The four observers all do measure different values for the
speed of sound. Those values are, as you would have guessed, very nearly
the sum of the official speed of sound and the speed of each vehicle. Why is
that? Because sound moves at its “official” speed—about 700 miles per
hour, or 340 meters per second—with respect to air, the medium in which
sound is a wave. Think back to Chapter 4, when I introduced the idea of
waves as disturbances of some medium. For sound waves that medium is air,
so to say “the speed of sound is 340 meters per second” has an unambiguous
meaning: it denotes the speed of sound waves relative to the air. If you’re
moving through the air, then you’ll get a different value for the speed of sound
relative to you—just as I suggested back in Figure 6.1. (By the way, the three
“moving” observers need to stick their measuring instruments outside their
vehicles to measure different speeds for sound because the air—the medium
for sound waves—within each vehicle shares that vehicle’s motion.)
But light doesn’t work that way because it requires no medium. Light
propagates through empty space, not through some material substance against
which you can gauge your own state of motion. Nineteenth-century physicists,
hung up on the concept of the ether, thought they had such a substance. If they
had been correct, then ether would have been for light what air is for sound,
and our four observers in Figure 8.1 would have measured different values
for the speed of light. The only one to measure the “right” value would be
that observer—if any—who happened to be at rest with respect to the ether.
But the nineteenth-century physicists were wrong. They tried to measure
differences in the speed of light associated with Earth’s motion through the
ether (that’s what the Michelson– Morley experiment was all about), but they
failed to find any difference whatsoever. Their failure led to seeming
contradictions—contradictions that vanished with Einstein’s abolition of the
ether and his assertion that the relativity principle holds for all of physics. So
light and sound do not behave analogously, and you can’t argue against
Figure 8.1 by analogy with sound.

Being Relativistically Correct

So what’s going on? How do the observers in Figure 8.1 all get the same
value for the speed of light? Although this follows from the relativity
principle, to say so isn’t a very satisfying explanation. There is an
explanation but it requires stretching your commonsense notions of time and
space.
One way out of our dilemma would be if something were wrong with
some of the observers’ measuring instruments. Maybe high-speed motion
does something strange to meter sticks and/or stopwatches, something that
just compensates for the motion and results in everyone measuring the same
value for c. Take that spaceship going by at half the speed of light, a speed
with which we have no first-hand experience. Might such rapid motion alter
the behavior of clocks and meter sticks in a way that we don’t notice in our
slower, everyday motions?
This is a tempting explanation, and there is, in fact, a grain of truth in it.
But in another sense it’s dead wrong, in a way that obviously and
dramatically violates the Principle of Relativity. Right here and now I want
to make sure we agree to avoid any language that even hints at violation of
that principle. So what’s wrong with saying that high-speed motion affects
measuring instruments? What’s wrong is the implication that some
instruments are moving at high speed and some aren’t—and that the ones that
aren’t moving are more “right.” That’s bad because it violates the essence of
the relativity principle, with its insistence that all observers in uniform
motion have equal claim to be “right” when they experiment with the
physical world. So it’s wrong—relativistically incorrect—to say something
like “The spaceship’s clock reads differently because it’s moving.” One is
often tempted to make such statements, and even some books on relativity
routinely do so. But such a statement is always misleading because it implies
that there’s a frame of reference—not the spaceship’s—that isn’t moving.
And it’s precisely the existence of such a frame, with its claim to be
absolutely at rest, that the relativity principle vociferously denies.
When talking about relativity, pledge to use only relativistically correct
language! Watch out for phrases like “the moving clock,” “the observer at
rest,” or “the high-speed spaceship.” Unless it’s clear what these things are
moving or at rest with respect to, then the phrases are meaningless. Worse,
they imply an absoluteness to motion and rest that is antithetical to the
relativity principle. Regarding Figure 8.1, we might say, “The stopwatch in
the spaceship reads differently from the one on Earth, because Earth and the
spaceship are in relative motion,” or, perfectly equivalently, “The stopwatch
on Earth reads differently from the one on the spaceship because Earth is
moving relative to the ship.” Those phrasings are consistent with the
relativity principle because they don’t mention motion in an absolute sense.
What isn’t consistent is a statement like “The stopwatch in the spaceship
reads differently from the one on Earth because the spaceship is moving,
period.”
By the way, you should be suspicious of words like “really” or “what
actually happens” when you’re talking relativity. Usually these have a hidden
implication; “really” might mean “from the viewpoint of someone on Earth”
and “what actually happens” might mean “as judged by someone who shares
my frame of reference.” That’s relativistically incorrect because of the
implicit assumption that some one point of view (frame of reference) has a
unique claim to judge reality or what actually happens.
Suppose you still insist you’re OK in claiming that the spaceship in
Figure 8.1 is moving, period. Then you haven’t yet accepted the truth of the
relativity principle. If you do accept that principle, then you have to give
observers on the ship exactly the same privilege: the spaceship’s astronauts
can claim that you, standing by the road, are the one who’s moving. So who’s
right? Both, or neither. Better to avoid the issue altogether and talk only about
relative motion. You: “The spaceship is moving relative to me.” Astronaut:
“You, and Earth, are moving relative to me.” Both statements are correct, and
neither implies any claim that anyone is at rest, or moving, in an absolute
sense. Only relative motion matters. That’s why it’s called relativity.
I’ve summarized this point about relative motion in Figure 8.2, which
shows two different stripped-down versions of Figure 8.1. Gone are the car
and the airplane, to avoid clutter. Gone, too, is the Earth—to help overcome
your natural prejudice that there’s something special about the solid ground
on which you, in Figure 8.1, are standing. Finally, gone is the traffic signal,
because—as double-star observations show and as application of the
relativity principle to electromagnetism makes clear—the speed of light does
not depend on the relative motion of the light source and the observer
measuring that speed. So there’s nothing special about your situation just
because you happen to be the one observer who’s at rest relative to the traffic
signal.
Figure 8.2a is drawn from your point of view, meaning that it shows the
situation in your frame of reference. Relative to this frame, you’re at rest and
the spaceship is moving to the right at half the speed of light. You obviously
measure c for the speed of the light. Your common sense might suggest that an
astronaut on the spaceship should measure 1.5c, but—hard as it may be to
accept—the Principle of Relativity tells you otherwise. Figure 8.2b shows
the situation in the ship’s frame. Relative to this frame, the ship is at rest, and
you’re moving to the left at half the speed of light. (Really? Remember, only
relative motion matters! Relative to an observer sitting at rest in the
spaceship, you move rapidly away to the left.) An astronaut on the ship
measures c for the speed of light. The astronaut’s common sense might
suggest that you should measure only 0.5c, since you’re moving away from
the light’s source, but relativity tells the astronaut otherwise. Both you and
the astronaut have equal claim to be in the right situation to do physical
experiments, and if you believe relativity then each of you has to grant the
other that same rightness.

Fig. 8.2 Take away the ground and the traffic signal, and it’s obvious that the roadside and rocket observers are in p erfectly
equivalent situations. Both think of themselves as being at rest, and both measure c for the sp eed of light. (a) From the roadside
observer’s viewp oint, the rocket is moving to the right at half the sp eed of light. (b) From the rocket’s viewp oint, the roadside
observer is moving to the left at half of c.
Space and Time are Relative

Let’s go back to Figure 8.1. How can all four observers measure the same
speed c for light? I mentioned that there’s a hint of correctness in the thought
that maybe their measuring instruments read different things. Notice I didn’t
say here that anyone’s instruments are wrong. This would imply that someone
else’s instruments are right, giving them the privileged position that the
relativity principle denies. Nor am I saying that one set of instruments
behaves differently from another; all work just the same way, timing the flash
of light over a fixed distance. And all give exactly the same result: the ratio
of the distance to the measured time is c, or 299 792 458 meters per second.
But how is this possible, given that the different sets of instruments are
moving relative to one another? It’s not that the instruments go “wrong,” but
something much deeper. It’s the nature of time and space. Time and space are
not absolute, but relative to one’s frame of reference. Measures of time
intervals and of spatial distances are simply different in different reference
frames. And, of course, no one has claim to the right measures; all reference
frames in uniform motion are perfectly acceptable for making valid
measurements of space and time. It’s the differences in space and time from
one reference frame to another that ensure that all four observers in Figure
8.1 measure precisely the same value for the speed of light. The spaceship
may be moving relative to you, but because its measures of space and time
are different from yours, you both come up with the same value c.

Sense and Common Sense

How can time be different for different frames of reference? That sounds
preposterous because your common sense suggests an absolute, universal
time ticking off equal-length seconds everywhere and in every state of
motion. It’s as if there could be a master clock somewhere that establishes
instantly a unique time throughout the whole Universe. Isaac Newton shared
your commonsense concept of universal time: “Absolute, true, and
mathematical time, of itself and from its own nature,” he said, “flows
equably without relation to anything external.”* But where is that absolute,
universal clock? In whose reference frame? Wouldn’t such a clock establish
a preferred state? And how would that sit with the relativity principle?
Your everyday experience may suggest that time is universal and
absolute, but is your experience broad enough to establish that absoluteness
beyond a doubt? No matter how sophisticated and well traveled you are,
you’ve probably never moved faster than the 600 miles per hour of a
commercial jet, relative to anything important to you. Even if you’re a
supersonic pilot or an astronaut, you’ve been limited to speeds of at most
about 5 miles per second relative to Earth. That’s still minuscule compared
with the roughly 186,000-miles-per-second speed of light. (The italicized
phrases are for relativistic correctness and also because you do move much
faster relative to things that you don’t directly perceive, such as that distant
galaxy we discussed in Chapter 2 or the electrons zooming through your TV
at 30 percent of c to create the picture on your TV screen.) So you lack the
experience to know for certain that high relative speeds don’t result in
different measures of space and time. And at the low relative speeds you
have experienced, maybe there are differences that are just too small for you
to notice.
Your commonsense concepts of time and space were established very
early in life. As a crawling baby, you explored space in its three dimensions.
Nothing led you to believe that crawling faster relative to your surroundings
in any way altered that space. Later, perhaps, came your notion of time. But
by the age when you were left at day care, started school, or whined “Are we
there yet?” from the back of the family car, you surely knew something about
the passage of time. Again, nothing in your experience from then until now
has suggested that the time between two events isn’t an absolute quantity, the
same for everyone.
But what if you had grown up crawling at 80 percent of the speed of light,
relative to your immediate surroundings? Then you’d have no need of this
book. Your commonsense understanding of time and space would be entirely
consistent with the Principle of Relativity, and what presented a stiff
intellectual challenge to no less than Albert Einstein would be intuitively
obvious to you. It’s just your provincialism—stuck on the hunk of rock called
Earth, limited to speeds far less than c relative to your planet—that makes
relativity hard to swallow. But relativity and your provincial common sense
are not actually in contradiction. Your limited experience may lead you to
think that time and space are absolute, the same for all who care to measure
them. But your experience doesn’t require that conclusion. It’s equally
consistent with space and time being relative to your frame of reference,
given that all the frames of reference you’ve ever occupied have moved so
slowly relative to one another that you don’t notice the difference. Slowly,
that is, in comparison with the speed of light.

Time Dilation

Exactly how do measures of time differ in different reference frames? Here

I’m going to argue that difference qualitatively and present without proof
how it works quantitatively. (Those who like math will find the details in the
Appendix.) First, though, I need to clarify an idea that’s already familiar to
you—the idea of an event. Events play a major role in relativity, because
they involve both time and space. An event is an occurrence, something that
happens at a specific place and time. Your birth is an event; it occurred
somewhere and at some time. Your reading this paragraph is another event;
it, too, is happening somewhere and at some time. The two events are not the
same, because they’re separated in time and probably in space as well.
(Although they could occur at the same place, at least in Earth’s reference
frame, if you happen to be reading in the very spot where you were born.) An
event is completely specified by giving a time and a place. For example, the
most recent major earthquake to strike the San Francisco Bay area resulted
from a specific event—a slippage along the San Andreas fault. The event’s
location (place) was at Loma Prieta, in the Santa Cruz Mountains, 60 miles
south of San Francisco and at a depth of 11 miles below Earth’s surface. Its
time was 5:04 PM PST on October 17, 1989. Another event, the landing of
the NEAR spacecraft, marked humankind’s first contact with an asteroid.
This event occurred at time 12:02 PM PST on Monday, February 12, 2001,
and its place was the asteroid Eros, then 196 million miles from Earth. Since
every event has a time and a place, it makes sense to talk about the time
intervals and spatial separations between any pair of events. For our two
events, the time interval is approximately 11.28 years and the spatial
separation is 196 million miles—as measured in a frame of reference at rest
with respect to Earth.
 I’m now going to convince you that the time interval between two events
cannot be the same for two observers in motion relative to each other. The
argument I’ll use is based solidly in the Principle of Relativity and its
consequence, the invariance of the speed of light. I’ll use a rather artificial-
seeming situation involving a bouncing light beam, but don’t come away
thinking this is only about light or about the particular situation I’ll describe.
Rather, it’s about the nature of time. The example I present serves to
illuminate what is, in fact, a universal aspect of time—namely, that measures
of the time between two events differ in reference frames that are in relative
motion.
So here’s the situation, shown first in Figure 8.3a. There’s a rectangular
box with a light source at one end and a mirror at the other. A brief flash of
light leaves the source. We’ll call this occurrence—the departure of the light
flash from the source—event A. The light travels up the box, hits the mirror,
reflects, and returns to the source. We’ll call event B the return of the light to
the source. That is, event B is completely specified by stating that it occurs at
the bottom of the box (B’s place) at the exact instant that the light flash
reaches that place (B’s time). The light’s round-trip takes some time, which
is the time between events A and B as measured in a frame of reference at
rest with respect to the box. We need that clarification because the time
between two events is not absolute but depends on one’s reference frame.
That’s just what I’m trying to demonstrate.
Fig. 8.3 A “light box,” used as a clock to measure the time interval between two events. Event A is a light flash leaving the
source. Event B is the light returning to the source, after reflecting off a mirror at the top of the box. Dashed line is the light
p ath. (a) The situation in a reference frame at rest with resp ect to the box. The light travels a total distance twice the box length
L. (b) The situation in a reference frame in which the box moves to the right at sp eed v. The light travels farther, but because it
has the same sp eed c, the time between the two events must be longer.

You could calculate the time between the two events quite easily if you
knew the length of the box. Let’s call that length L. Then the light goes a total
distance of twice L, because it makes a round-trip. And how fast does the
light go? At speed c, as always. Knowing the distance and speed, it’s simple
to calculate the time—and I do that explicitly in the Appendix. But for this
qualitative argument, all you need to know is that the light makes a round-trip
of twice the box length and that it does so at speed c.
I warned you earlier to be suspicious of pictures in relativity books.
Every picture is drawn from a particular point of view, that is, from the
perspective of a particular reference frame. To understand what’s going on,
you need to know exactly what frame that is. In the case of Figure 8.3a, the
reference frame of the picture is at rest with respect to the box, so it’s in that
reference frame that we’ve been discussing the time between events A and B.
Now let’s look at the situation from another reference frame, one moving
relative to the box. Equivalently, the box is moving relative to this new
reference frame. So suppose the box is moving to the right, at some speed v,
relative to the new frame (v here is for “velocity”). Figure 8.3b shows the
situation in this reference frame. Because the box is moving relative to the
reference frame, I’ve needed to show it in different locations at several
different times. In particular, it’s shown at the time of event A, when the light
flash leaves the source; at the time of event B, when the light returns; and at
an intermediate time when the light reflects off the mirror at the end of the
box.
We want to know the time between events A and B as measured in the
reference frame of Figure 8.3b. To find that time we need the distance the
light travels between its departure from and its return to the source, and we
need to know how fast it goes. Without doing any math, you can see that the
light’s path in Figure 8.3b is longer than it is in Figure 8.3a, which was
drawn from the perspective of a reference frame where the box was at rest.
That’s because the light, relative to the reference frame of Figure 8.3b, takes
a diagonal path heading to the mirror and another diagonal path coming back.
Those diagonals are each necessarily longer than the length L of the box,
since they incorporate both that length and some motion of the light sideways,
along the direction of the box’s relative motion. Again, there’s nothing fishy
about this. If I’m in a bus and I throw a ball straight up, it goes straight up and
down relative to me. But if you’re standing by the road you see the ball take a
curved path that’s longer than the path it takes in a reference frame at rest
with respect to the bus. Similarly, the light takes a longer path in a reference
frame relative to which the box is moving.
The only other thing we need to get the time between the light’s departure
and return—that is, the time between events A and B—is the light’s speed.
Here’s where the relativity principle comes in. Again, one consequence of
that principle is the invariance of the speed of light. So the light goes at c
relative to the reference frame of Figure 8.3b. But it goes farther in this
reference frame than it did in the frame of Figure 8.3a, at rest with respect to
the box. Since it has the same speed, c, in both frames, it must take a longer
time in the reference frame with the longer path—that is, in the reference
frame relative to which the box is moving. So we’re forced to conclude that
the time between events A and B in a reference frame in which the box is at
rest is shorter than the time between the same two events in a reference
frame with respect to which the box is moving.
Let me be clear that I’m not saying simply that the light took two different
trips, one of them longer, and therefore that the longer trip took a longer time.
That would be obvious without relativity. Rather, the same light took one and
the same trip, and we examined that trip from the viewpoints of two different
reference frames. The beginning and end of that trip are the events A and B,
and observers in both frames agree about what those events are and that they
indeed mark the endpoints of the light’s trip. What they disagree about is the
time between those same events.
That’s all well and good, you might say, in this highly artificial situation
where the two events in question happen to involve a light beam that must go
at c in any reference frame. Surely, you say, this has nothing to do with
everyday events of the sort you and I experience. But it has everything to do
with such events. In principle, I could have set up a “light box” at any event
you care to name—for example, at your birth—and arranged for the flash to
depart at the instant of that event. By moving the light box at the right speed I
could arrange for it to arrive at a later event of your choosing—for example,
your reading this chapter. And by choosing the length of the box (it would
have to be very long for your birth and reading events!), I could arrange for
the light to return to the source just as the box arrived at where you are as you
read this chapter. All the arguments I made from Figure 8.3 would apply
here, and we’re forced to conclude that the time between the events of your
birth and your reading this chapter is different in different reference frames.
This is about time, not about light beams and mirrors. The light box serves as
a device for exploring the nature of time and for concluding that the time
between two events is relative to one’s frame of reference. But the light box
doesn’t cause that difference; the difference is intrinsic in the nature of time.
Measurement of the time interval between two events with ordinary clocks
would, in principle, reveal exactly the same discrepancies between different
reference frames, regardless of the presence or absence of the light-box
device. In fact, all good clocks in a given reference frame will measure the
same time between two events, and that time will be different from that
measured by identical clocks in another reference frame in relative motion.
By a “good clock,” I mean any device that accurately measures the passage
of time. That includes the light box in Figure 8.3, an atomic clock, your
bedside alarm clock, or your wristwatch. It also means time-dependent
processes like the vibration of a radio wave, the beating of your heart, or
even the biological clocks that govern the aging of your body. All measure
the same underlying phenomenon—time itself.
Before exploring further this strange new understanding of time, let’s tie
up a few loose ends. First, be very clear where and how the Principle of
Relativity entered my argument. It entered in the assumption that the speed of
light was c in both reference frames. Again, that assumption follows directly
from the relativity principle as applied to electromagnetism. Had we not
made that assumption, you could have argued that the light in Figure 8.3b was
going faster than in Figure 8.3a because it shared the box’s horizontal motion.
Then, even though it was going farther, it would have taken the same time in
both frames. But you can’t make that argument for the same reason that you
can’t argue for different speeds of light for the observers in Figure 8.1. It’s
the Principle of Relativity that rules here, not some commonsense but
incorrect notions you’ve developed about how light ought to behave. And the
relativity principle implies that the speed of light is the same in all uniformly
moving reference frames.
 There’s a more subtle point you might want to challenge. In arguing that
the light’s path in Figure 8.3b is longer than in Figure 8.3a, I’ve implicitly
assumed that the box length L is the same in both reference frames. But I’ve
mentioned repeatedly that relativity affects both time and space. So how do I
know the box length isn’t different in a way that shortens the path in Figure
8.3b and thus voids the time difference? It is true that measures of space, too,
depend on reference frame (we’ll soon see just how) but that only happens
for measures taken along the direction of relative motion between two
frames. There’s no difference for measures, like that of the box length L in
Figure 8.3, that are perpendicular to the direction of relative motion. You
don’t have to take this on faith, because it follows directly from the relativity
principle. However, I’d like to hold that argument until we’ve explored
further our new discovery about the relativity of time.

“Moving Clocks Run Slow”

The shortening of the time interval between our events A and B in the light-
box frame is known as time dilation. It’s as if clocks—and indeed, all
manifestations of time—are running more slowly in the light-box reference
frame than they are in the frame relative to which the box is moving. Many
books on relativity summarize time dilation with the phrase “moving clocks
run slow.” This is a dreadful phrase, as relativistically incorrect as they
come! (It’s also grammatically incorrect; the last word should be the adverb
“slowly.”) So what’s wrong with “moving clocks run slow”? Here’s what:
Who’s to say that one clock is moving and another isn’t? Given the relativity
principle, the phrase “moving clocks” is utterly meaningless and worse,
since it applies an absoluteness to motion that is antithetical to the very
essence of relativity.
Here’s a better, albeit wordier, description of time dilation:

The time between two events is shortest when measured in a

reference frame where the two events occur at the same place.

This description is fully consistent with the light-box example that led us to
time dilation. In the light-box reference frame, events A and B do occur at the
same place, namely the bottom of the box. The same place? Even though the
box is moving? If you’re thinking that way, then you’re not yet fully accepting
relativity. The box’s reference frame is a perfectly good one for doing
physics, and no one can make an absolute claim that the box is moving. To
someone in the box, it’s at rest. An observer sitting in the bottom of the box
when the light flash goes off (event A) can stay in the same place and will
later be present when the flash returns (event B). Since the observer doesn’t
have to move relative to the box, the two events occur, for this observer, at
the same place. If you still want to claim that the two events aren’t “really” at
the same place because the box is “really” moving, then you’re granting a
special status to the other reference frame, the one relative to which the box
is moving. If you insist on doing so, then you’re in violation of the Principle
of Relativity. Watch out for that word “really”!
You might also object by citing a more concrete example of two events,
namely your birth and your reading this chapter. Suppose you were born in
San Francisco and you’re now sitting in New York. Am I saying that San
Francisco and New York are in the same place? Of course not. But to an
observer sitting in a car that left San Francisco at your birth and just now
arriving in New York after a slow but steady trip, the events of your birth and
of your reading occur at the same place, namely right at the car. Note the
distinction here between events and places. San Francisco and New York are
indeed in different places, and different observers agree about that (although,
as you’ll soon see, they disagree about how far apart they are). But as
measured by an observer in a reference frame moving with respect to Earth,
the positions of the two cities are continually changing. By choosing just the
right motion relative to Earth, our carbound observer can arrange for those
positions to be the same—namely, right at the car—at both your birth and
your reading of this chapter. That’s what I mean when I say that two events
occur at the same place in some reference frame.
Figure 8.4 illustrates the meaning of my description of the time-dilation
phenomenon. The picture is drawn in the reference frame of two clocks, C1
and C2, that are located at different places. A third clock, C, moves relative
to the others in such a way that it passes first C1 and then C2. We’ll assume
that C passes very close to the other clocks. Then just as C passes C1 we
have a distinct event, occurring at a definite place and a definite time. We’ll
call that event A. Event B will be C passing C2. In the reference frame of
clocks C1 and C2, the two events obviously occur at different places,
namely, the distinct locations of the two clocks. But clock C is moving
relative to the others in such a way that it’s present at both events. That is,
events A and B occur at the same place in C’s reference frame—in just the
same way that your birth and reading occurred at the same place in the
reference frame of the car in the previous paragraph. Then, says time
dilation, the time interval between events A and B will be shorter as
measured on clock C than as measured by clocks C1 and C2. I’ve indicated
this time difference on the clocks in Figure 8.4.

Fig. 8.4 Two clocks, C1 and C2, are at rest with resp ect to each other, and the figure is drawn in the reference frame of these
two clocks. A third clock, C, moves relative to C1 and C2 in a direction that takes it first p ast C1, then on p ast C2. Clocks C
and C1 read the same time when the two coincide; this is event A. When C reaches C2 (event B), C reads less elap sed time
than C2. Thus, the time between events A and B is shorter in C’s frame of reference. (Here C’s sp eed is such that its time is
half that measured by C1 and C2.)

There’s an important sense in which the situations in the two reference

frames involved in our time-dilation examples are not equivalent. In one
frame—the frame of the light box in Figure 8.3 or of clock C in Figure 8.4—
it takes only one clock to measure the time interval between the events in
question. But in the other frame it necessarily requires two clocks. That’s
because the events occur at the same place in the first frame but at different
places in the second. (You might ask why you can’t use a single clock in
place of clocks C1 and C2, and watch for the occurrence of the two events
from afar. You could do that, but then you’d have to compensate for the travel
time of light coming from the two events to your observing location. You
already know enough relativity to realize that such compensation is going to
require some care. In any event, an observer right at clock C doesn’t need to
apply any such compensation, so the two situations still wouldn’t be
equivalent.)
There’s another issue in using two separated clocks to measure the time
interval between two events. Obviously, those two clocks have to be
synchronized—meaning they both read the same time at the same instant.
That’s not hard to achieve, but it does require a little thought. When clock C1
reads exactly noon, for example, it could send out a light flash. When the
light flash reaches clock C2, an observer there could set C2 to read noon
plus the time it took the light to travel from C1 to C2. Then the clocks
would be synchronized. Now, an observer looking at the two clocks
wouldn’t see them reading exactly the same time unless the observer were
equal distances from both clocks. An observer closer to one clock would see
that one reading a later time, since light from that clock would have left more
recently because of the shorter distance. But knowing the distances to the two
clocks that observer could nevertheless infer that they were, in fact,
synchronized. The observer would thus observe that the clocks were
synchronized even though she doesn’t see them reading the same time.
Note here the important distinction between seeing and observing.
Seeing, when what you see involves objects at different distances, doesn’t
give an accurate picture of what is, because of the different travel times for
light from the two objects. (When I say “is” here, I mean what “is” for an
observer in your reference frame; I’m not referring to some absolute truth
that’s independent of reference frame. Thus clocks C1 and C2 are
synchronized to an observer at rest with respect to the two clocks, even if
that observer doesn’t see them reading the same time.) When I say that
someone observes a given situation, I’ll always mean that that person
determines what actually happens as judged from his or her frame of
reference. Often such observation means more than just looking; instead, it
means looking and then compensating for light’s travel time.
You might think I’m making a big fuss over not much, because light
travels so fast that the differences in travel time from different objects are
negligible. That’s true for most everyday situations. But when objects move
relative to you at very high speeds, or when they’re very far away, the
differences become significant. Those are precisely the situations in which
relativistic effects are evident and in which your commonsense notions of
time and space become inadequate. As a concrete example, consider
astronomers trying to take a snapshot of galaxies in a particular region of the
sky. The nearest galaxies might be so close that their light takes only a few
million years to reach us. For the most distant, that time becomes billions of
years. So the astronomers’ image is not a snapshot at all, in that it doesn’t
show the galaxies as they are, or even as they all were at some common time
in the past, but rather as they were at times that vary dramatically from galaxy
to galaxy. What the astronomers see is not what is.

Getting Quantitative
Just how significant is the time difference between events as measured in
different reference frames? That depends on the speed of the frames’ relative
motion. For relative speeds small compared with the speed of light, time
dilation is not at all obvious and is very difficult to measure. For relative
speeds approaching c, though, the effect becomes dramatic.
Although I’m taking a nonmathematical approach to relativity in this
book, I nevertheless want to show you quantitatively the effect of time
dilation. One simple formula sums up that effect. If you like math, you can
follow its derivation in the Appendix. Even if you aren’t into math, you’ll
recall that sometime in high school or earlier you learned how to use the
Pythagorean theorem to find the diagonal of a right triangle. That involved
taking the sum of the squares of the two shorter sides, then taking the square
root. What’s this got to do with time dilation? Simply this: the light path in
Figure 8.3b forms the diagonals of two right triangles. As the Appendix
shows rigorously, finding the light travel time in the reference frame of
Figure 8.3b thus involves the Pythagorean theorem. So it’s no surprise that
the formula for time dilation involves squares and square roots:
 I’ve labeled this time-dilation formula to make it clear exactly what each
term means. The formula gives the time, t' (read “t prime”), between two
events A and B, as measured in the reference frame in which two events
occur at the same place (e.g., the frame of clock C in Figure 8.4). To the right
of the equal sign, t is the time as measured in a reference frame in which the
two events occur at different places (e.g., the frame of clocks C1 and C2 in
Figure 8.4). The symbol v under the square-root stands for the relative speed
of the two reference frames, given as a fraction of the speed of light. That is,
v = 0.5 means half the speed of light, v = 0.9 means 90 percent of the speed
of light, and so forth. So what the formula says is that the time t' in the frame
where the events occur at the same place is given by the time in the other
frame, multiplied by the square root of 1 −v2. Let’s take a closer look at this
quantity. If v = 0, then the two frames aren’t in relative motion and we’re left
with the square root of 1, or just 1. So our formula gives t' = t. Of course: In
this case the two frames are really one and the same reference frame, and the
formula is just telling us that observers in the same reference frame all agree
on the time between events. But if the two frames are in relative motion, then
the two times are no longer equal. Consider first a fairly low relative speed,
say 10 percent of the speed of light, so v = 0.1. Then v2= 0.01, and the
quantity √ 1− v2 becomes the square root of 0.99, which is very nearly 1. So
even in this case—a relative speed of nearly 19,000 miles per second—the
two times are still very close. It’s only when the relative speed becomes a
significant fraction of c that time dilation becomes substantial. With v = 0.8
for the relative speed between two frames, for example, you can convince
yourself that √ 1− v2 = 0.6, meaning that the time between two events
measured in a frame where the events occur at the same place is only a little
over half what it is in the other frame. At v = 0.99, the square root is about
0.14, and the one time is only about a seventh of the other. Finally, you might
wonder about the case v = 1, corresponding to relative motion at the speed of
light. Here the formula gives t' = 0. This suggests that time does not pass at
all in the frame where the two events occur at the same place! But as we’ll
see in Chapter 12, a relative speed of c is not possible for frames of
reference associated with physical objects like human observers or clocks of
any kind.

Wrapping It Up
The crucial point of this chapter is that measures of the time between events
need not be the same in two reference frames in relative motion. In
particular, observers in two different reference frames measure different time
intervals between the same two events—with the shortest time measured by
an observer for whom the two events occur at the same place. Although I
spent a long time elaborating on this point, it follows directly from the
Principle of Relativity and its consequence, the invariance of the speed of
light. If you accept the Principle of Relativity, then you can’t logically escape
this conclusion about the relativity of time.
But is this strange new behavior of time, as embodied quantitatively in
the time-dilation formula, at all relevant to anything? Can we find or imagine
situations in which it occurs and is important? Or is all this just an academic
exercise?
As the numerical examples in the preceding section show, we can expect
obvious time-dilation effects only when relative speeds are very high—close
to the speed of light. We might detect time dilation at lower relative speeds,
but then only with very sensitive experiments. At everyday relative speeds,
including those of jet aircraft and even today’s spaceflight, time dilation is
simply too small for us to notice directly.
Even so, time dilation does occur and is measurable. The effect shows up
dramatically in experiments involving subatomic particles moving, relative
to Earth, at speeds approaching c. As I outlined in Chapter 1, it’s even been
measured in clocks flown around Earth in ordinary aircraft—although here
the effect is minuscule. We can imagine a future with high-speed space travel,
where observers on Earth and passengers in spacecraft would measure very
different times between the same events. We’ll explore these realities and
possibilities in the next chapter.

*Isaac Newton, Princip ia (Berkeley, CA: University of California Press, 1934); originally p ublished in London in 1687 and
trans. by Andrew M otte in 1729. This p hrase ap p ears near the beginning of the work, in the “Scholium,” a discourse on time,
sp ace, and absolute versus relative motion that follows Newton’s definitions of basic p hy sical quantities.
CHAPTER 9

STAR TRIPS AND SQUEEZED SPACE

• • •

Suppose we have a spaceship capable of traveling, relative to Earth, at 80

percent of the speed of light. We’ll dismiss the obvious technical challenges
in building such a craft, and we’ll also assume that it can jump essentially
instantaneously from being at rest on Earth to its cruising speed of 0.8c. Such
an acceleration would surely be fatal, but we’ll imagine we’ve somehow
overcome this problem. Finally, we’ll equip the ship with the latest
instrumentation, including an accurate clock just like those used on Earth.

Years and Light-Years

Before we set out on a star trip, I want to introduce a way of describing

interstellar distances that makes a lot more sense than feet and miles, or
meters and kilometers. Instead of those familiar units, we’ll measure
distances in light-years. Despite the word “year” in its name, the light-year is
a unit of distance, not time. One light-year is, simply, the distance light
travels in a year. How far is that? It doesn’t really matter, but you can find out
approximately, if you want, by multiplying 186,000 miles per second by the
number of seconds in a year (about 30 million). The result is about 6 trillion
miles. It’s more meaningful to compare the light-year with real distances. For
example, the nearest stars beyond our Solar System are a few light-years
away. Our Milky Way galaxy measures about 100,000 light-years across, and
we’re about 30,000 light-years from its center. The Milky Way’s nearest
large neighbor, the great spiral galaxy in the constellation Andromeda, and
the most distant object visible with the naked eye, is about 2 million light-
years away. The most remote objects astronomers have spotted with the
Hubble Space Telescope are quasars some 12 billion light-years distant. So
the light-year is a convenient unit when we’re talking interstellar or
intergalactic distances. But it’s too big to be useful within our own Solar
System; the Sun, for example, is only about 8 light-minutes from Earth, and
the Moon is just over 1 light-second away. Either of these distances is only a
tiny fraction of a light-year.
In describing travel in our hypothetical spaceship, then, we’ll use the
light-year as our unit of distance and we’ll use the year as our measure of
time. Finally, it would be nice to express the all-important speed of light in
these new units. That’s easy! Because a light-year is defined as the distance
light travels in 1 year, the speed of light is, by definition, 1 light-year per
year. So our 0.8c spacecraft goes 0.8 light-years per year relative to Earth.
Its speed, as measured in light-years per year, is also precisely the number v
—the speed as a fraction of the speed of light—that we’re supposed to use in
our time-dilation formula. So working in light-years and years is going to
make things easier.

Star Trip! One-Way

Suppose you set out in that 0.8c spaceship to visit a star 20 light-years from
Earth—as measured in Earth’s frame of reference. We’ll assume Earth and
star are essentially at rest relative to each other (a reasonable assumption,
given that nearby stars don’t move very fast relative to Earth). Then Earth
and star are in the same reference frame, which I’ll call the Earth–star frame.
We’ll suppose that there are identical clocks located on Earth, at the star, and
in the spaceship, and that the Earth and star clocks are synchronized as
described in the preceding chapter. So the Earth and star clocks are like the
clocks C1 and C2 in Figure 8.4, and the ship clock is like clock C in that
figure.
How long does the star trip take? For the Earth–star frame the answer is
easy: We know the Earth–star distance (20 light-years) and we know the
speed. Back in elementary school you learned that distance = speed × time;
after all, that’s just what speed means, namely, how far you go in a given
time. Therefore time = distance ÷ speed. So the trip time as measured in the
Earth–star frame is

t = (20 light-years) ÷ (0.8 light-years/year) = 25 years.

Does this make sense? Sure: If the ship were going at the speed of light (1
light-year per year) it would take exactly 20 years to make the 20 light-year
journey. It’s going a little slower, so the trip should take a little longer. If all
the clocks read 0 years when the ship starts out, then our answer for t shows
that the Earth and star clocks will read 25 years when the ship reaches the
star.
What about the ship clock? Like clock C in Figure 8.4, this is a clock for
which the two events of interest—the ship’s departure from Earth and its
arrival at the star—occur at the same place. So it’s going to take less time as
measured on the ship clock than it does on the Earth and star clocks—a
conclusion forced on us by the argument of the previous chapter, an argument
based on nothing more than the Principle of Relativity. The time-dilation
formula we developed in the previous chapter applies to just such a clock as
our ship clock, so we can work out the time t' as measured on the ship:

t’ = t × √1 − v2 = 25 years × √1 − 0.82 = 25 years × √ 1 − 0.64

= 25 years × √0.36 = 25 years × 0.6 = 15 years.

(You could do the math on a calculator, but I showed all the details because
with v = 0.8 the square root works out so nicely to give 0.6 for the time-
dilation factor.) Figure 9.1 summarizes your star trip and the different times
involved.
The difference here is substantial! In the Earth–star frame, the time
between the ship’s departure and arrival is 25 years. In the ship’s frame, it’s
only 15 years—for a difference of 10 full years. Like everything that’s come
before, this conclusion follows inescapably from the Principle of Relativity.
But what does it mean, besides that the ship’s clock seems to “run slow”
compared with the Earth–star clocks? In particular, what happens to you as
you travel on the ship? Does the trip seem to take just 15 years, or is it
“really” a 25-year trip in a spaceship whose clock somehow isn’t reading
right? If you’re thinking the latter, then you’re being relativistically incorrect!
The ship’s frame is just as good as the Earth–star frame for exploring
physical reality. What happens in the ship’s frame is every bit as real and
valid as what happens on Earth. It isn’t just that the ship’s clock reads 15
years—it’s that, in the ship’s frame of reference, only 15 years of time have
elapsed between departure from Earth and arrival at the star. All
manifestations of time reflect this interval. In particular, you arrive at the star
15 years older than you were when you left Earth—not the 25 years older
that your earthbound friends might expect.

Fig. 9.1 At 0.8c, the trip from Earth to star takes 25 y ears in the Earth–star reference frame but only 15 y ears in the
sp aceship ’s frame. Times are shown on three digital clocks, one on the ship (shown twice) and one each on Earth and star.
Figure, including clock readings, is a comp osite of two different times: first when the ship p asses Earthand then when it p asses
the star.

Would you be aware, while traveling, that you’re aging more slowly than
usual? Would you feel your heartbeat slow down and notice your hair
growing more slowly? Absolutely not! To you, in your spaceship, everything
seems perfectly normal. It must! Why? Because you’re in a perfectly good
frame of reference for exploring physical reality. The laws of physics work
just as well for you as they do on Earth. If you felt something strange because
of the ship’s motion, something you didn’t feel on Earth, then you could
genuinely say that there was something unusual about the spaceship’s
reference frame, something that wasn’t true of the Earth’s frame—namely,
that the ship was moving in some absolute sense while Earth wasn’t. But, of
course, it’s just such specialness of one reference frame over another that
relativity denies. So everything—including the passage of time in all its
manifestations from atoms to clocks to human bodies—seems perfectly
normal onboard the spaceship.
 At this point you’re probably full of questions. What if you radio back to
Earth periodically, maybe at each birthday? Will your friends back home
recognize that you’re aging more slowly than they? Suppose some friends
your own age had traveled earlier to the star, via a much slower spaceship.
When you stop at the star and join them, will you really be younger? Just
what is it that’s holding back time on the spaceship? What’s getting into every
clock, into the aging mechanism of each cell in your body, and into every
electron whirling about every atom, to slow them all down?
I’ll answer all these questions eventually, but for now I want to address
just the last one. The answer is that nothing is holding things back. Don’t go
looking for a mechanism that slows everything down. To do so is to say that
there’s really a universal time but that something unusual happens in the
moving spaceship to make all time-consuming processes take longer. That’s
no good because, again, it singles out Earth’s frame as the one in which this
time-slowing mechanism isn’t at work, and it identifies the ship’s frame as
one in which something unusual happens because the ship is moving. To say
those things about Earth and spaceship is to violate the relativity principle.
There’s one subtle point I need to make. We’ve applied the idea of time
dilation—that the time between the same two events is different when
measured in reference frames in relative motion—to the duration of your star
trip. We can do that if both reference frames, namely the Earth–star frame and
the ship frame, are indeed in uniform motion. That’s because the Principle of
Relativity says that the laws of physics are equally valid in all reference
frames in uniform motion. If the ship starts from rest relative to Earth, then
travels to the star and slams on its brakes, it’s certainly not moving uniformly
while it starts and stops. The reason I’ve assumed the ship can accelerate
instantaneously to 0.8c relative to Earth, and then stop abruptly at the star, is
so that it spends essentially the entire journey in uniform motion. To be
completely correct, we should identify the departure event as occurring just
after the ship has accelerated. Since the acceleration takes place
instantaneously, the ship is still right at Earth but it’s now moving relative to
Earth. Its clock still reads 0 because the acceleration took essentially no
time. The ship then moves, uniformly, relative to Earth and star, until the
arrival event, when it’s at the star but hasn’t yet applied its brakes. Between
the departure and arrival events so defined, the ship indeed moves uniformly,
and we can apply the Principle of Relativity and its logical consequence, the
time-dilation formula, to reach our conclusion that the time between
departure and arrival is 10 years less for those riding the ship than for those
remaining on Earth or at the star.
A cleaner way to present the star trip would have been to imagine an
alien being in a spacecraft that happens to come whizzing by Earth at a steady
0.8c. Later the alien passes a star that, as measured in the Earth–star frame, is
20 light-years from Earth. We can draw exactly the same conclusion in this
case as we did before, namely, that the alien measures 15 years between
passing Earth and passing the star, while observers with synchronized clocks
at Earth and star report a 25-year interval between the time those on Earth
see the craft go by and when observers at the star see it pass. What’s simpler
about this situation is that we don’t have to think about the alien spacecraft
starting and stopping; it’s truly in uniform motion relative to Earth the whole
time, so it’s completely obvious that the relativity principle applies. But I
emphasize again that it also applies to your star trip, at least during the
interval when you’re moving uniformly relative to Earth. Later in this chapter
we’ll reconsider the effects of starting and stopping.

Squeezing Space

How can the spaceship get from Earth to star in only 15 years? After all, the
distance is 20 light-years, so even light should take 20 years. As you’re
probably aware, and as I’ll elaborate on in Chapter 12, no material object
can go faster than light. So how can the spaceship get to the star in only 15
years?
The answer lies in the fact that measures of space, as well as of time, are
different in different reference frames. Let’s look at things from your
viewpoint as you ride the spaceship. You know you’re moving at 0.8c
relative to Earth and star, and you know that your journey takes 15 years. So
how far have you gone? You’re in a perfectly good frame of reference for
doing physics, so the elementary-school formula distance = speed × time
works just fine for you. So to you, the distance you’ve traveled from Earth to
star must be

d' = (0.8 light-years/year) × (15 years) = 12 light-years.

(I’ve called this distance d' [“d prime”] for consistency with my earlier
notation, where I used t' [“t prime”] to represent time measured in the ship
frame.)
There’s no inconsistency here and no faster-than-light travel. To you in
the spaceship, the Earth–star distance is 12 light-years, and at 0.8 light-years
per year, it makes perfect sense that your trip takes a little over 12 years—15
years, to be precise. What’s troubling is that something you thought was
objectively real and absolute, namely the distance between two objects,
simply isn’t. Like measures of time, measures of space also depend on one’s
frame of reference.
Note, by the way, that the 12-light-year distance in the ship frame is
precisely 60 percent of the 20-light-year value that the Earth–star distance
has in the Earth–star frame. That 60 percent, or 0.6, is just the relativistic
factor that we calculated for time dilation with v = 0.8c. In fact, that’s
generally true. If the distance between two objects is d in a frame of
reference where the two are at rest, then in a reference frame moving at
speed v relative to the objects, the distance will be contracted by this same
factor, giving

d' = d × .

For this length contraction to occur, the relative motion must be along a line
between the two objects. If the motion is perpendicular to that line, then
there’s no contraction; if it’s at an angle, then the contraction factor is
somewhere between.
I referred to the shrinking of the distance between Earth and star, or any
to other objects at rest with respect to each other, as length contraction.
That’s because the two objects could be the opposite ends of a single
physical thing, like a ruler. To someone moving relative to the ruler, it’s
contracted to a shorter length. The ruler may still say “12 inches,” but as
measured by an observer moving relative to the ruler, it won’t measure 12
inches. But isn’t the ruler really 12 inches long? To ask that is to suppose
there’s one special frame in which measures of length or distance are correct
and that therefore they’re wrong in other frames. Obviously, that view
violates the relativity principle. On the other hand, one’s relationship with an
object like a ruler is certainly simplest in a frame of reference at rest with
respect to the object. For that reason the length measured when one is at rest
with respect to an object is called the object’s proper length. Here “proper”
doesn’t mean “correct” as much as it does “proprietary”—in the sense of
“belonging” to the object, or intrinsic to the object in its own frame of
reference. But that doesn’t mean the object’s proper length has any
transcending reality. Observers in different reference frames will measure
different lengths for an object, and they’ll all be correct. Measures of space
and time just aren’t absolute. An object is longest in a reference frame where
it’s at rest and shorter in any other frame. Just how short depends on one’s
speed relative to the object, as given by the formula above.
Once again I need to remind you of the difference between observing and
seeing. When you observe an object moving relative to you, you measure its
length by, perhaps, noting where its front and back ends are in relation to a
ruler or meter stick at rest in your reference frame. Since the object is
moving, this can be a bit tricky. You have to be careful to note the positions
of the front and back ends at the same time. But you can do that, and the
result will be the contracted length of the object as measured in your
reference frame. However, that doesn’t mean you’ll see the object
contracted. That’s because light from different parts of the object reaches you
at different times, and those time differences are significant for an object
moving fast enough relative to you that its length is noticeably contracted.
Remarkably, the object appears not contracted but rotated! However, I’m not
going into the details of how this comes about because it’s a bit of a
distraction from the bigger theme of the nature of space and time.
Finally, a historical note. The length-contraction formula d' = d × is
precisely the remedy Lorentz and Fitzgerald proposed in the late 1800s to
resolve the quandary of the Michelson–Morley experiment. If the Michelson–
Morley apparatus contracted by this amount in the direction of its motion
through the ether, Lorentz and Fitzgerald correctly argued, then the travel
times for light along the two legs of the apparatus would remain the same,
and the experiment wouldn’t be able to detect Earth’s motion through the
ether. So Lorentz and Fitzgerald got it partly right, in that they correctly
predicted a motion-induced contraction of material objects. But they
remained philosophically mired in a relativistically incorrect way of
thinking, because for them the contraction occurred against a background of
absolute space and time. Theirs was a contraction of material objects in an
uncontracted space. The relativistically correct interpretation of length
contraction is that measures of space itself differ in different reference
frames and that differing measures for the length of material objects reflect
this underlying relativity of space.

It Really Happens!

Does all this really happen? Does time really pass more slowly on your
spaceship than it does on Earth and star? Is the Earth–star distance really less
for you as you ride the spaceship? The answer to these questions is yes—
although if you’re in a deeply relativistic mode of thinking then these
questions themselves may be bothersome. Why, after all, can’t you on the
spaceship consider Earth to be moving, and therefore its clock to “run
slow”? You can, and I’ll return to this point later. For now, though, let’s
consider how we might show that it really happens. Unfortunately, we don’t
yet have the technology to carry out the star trip, so we can’t directly
experience time dilation and length contraction. We do have access to objects
that move relative to us at very high speeds but those objects are too small
for us to ride on, because they’re subatomic particles. Surprisingly, though,
they carry “clocks” that allow us to confirm quite dramatically the effects of
time dilation.
Here I’m going to describe an experiment that was done in the 1960s
with the express purpose of showing directly that time dilation and length
contraction do occur.* Special relativity had been thoroughly verified long
before the 1960s, so this experiment was designed and performed especially
to be convincing to folks other than scientists. I’m going to introduce the real
experiment first by imagining a fictitious but analogous experiment closely
related to our Earth–star trip.
Imagine I have some unusual clocks with the remarkable property that
they self-destruct in 20 years, exploding into a jumble of hands and gears.
These aren’t very accurate clocks, so some explode after only 18 actual
years, a very few as early as 12 or 15 years, while a few last 25 or 30 years.
If I take 100 new clocks and wait 20 years, roughly half of them will be gone.
If I wait 25 years, only a very few will be left. But if I wait only 15 years,
nearly all of them will still be around.
 Let’s load your spaceship with 100 of these clocks, and send you off on
that star trip at 0.8c. Your friends remain on Earth, although a few of them
have already traveled—again, on a much slower spaceship—and are waiting
at the star. When the ship reaches the star, what will you or your star-based
friends find? Will the spaceship be full of clocks or will most of them have
self-destructed before the journey is over? Your Earth–star friends, who
agree that your trip takes 25 years, might expect to find very few clocks. But
for you the trip takes only 15 years, so nearly all the clocks should still be
around. So which is it? This is something we can test objectively by opening
up the ship when it reaches the star and looking for the clocks. If we don’t
find many, then we can conclude that the trip time was really 25 years, even
for the traveling clocks. If we do find a lot of clocks remaining, then we’re
forced to conclude that the trip time as judged by these clocks was
considerably less than the 20 years it takes before half the clocks will
explode.
If we did this experiment, which of course is no more possible than the
original star trip with a single accurate clock onboard the ship, then relativity
says that we should indeed find lots of clocks, because in the ship frame the
trip time is only 15 years.
In fact, an almost identical experiment is possible, and it’s the one that
was done in the 1960s. Instead of clocks, the experiment uses subatomic
particles that are radioactive—meaning that they self-destruct by exploding
into several other particles. The particular particles are called muons, and
they’re formed when cosmic rays slam into atoms high in Earth’s atmosphere.
A steady rain of these muons comes downward through the atmosphere at
speeds approaching c. The experiment consists of two parts. First,
experimenters high on a mountaintop (Mount Washington, New Hampshire)
use a muon detector to determine the intensity of the “muon rain” (Figure
9.2a). The detector is designed to catch only muons moving with a narrow
range of speeds very close to 0.995c. The experimenters detect just over 500
such muons each hour at the mountaintop. Now the time it takes muons to
self-destruct is well known, although the 1960s experimenters actually
measure it again as part of their experiment. That self-destruct time
(analogous to the 20 years for our self-destructing clocks) is short enough
that even muons moving at 0.995c would seem very unlikely to make it from
the altitude of Mount Washington down to sea level without self-destructing
on the way.

Fig. 9.2 M uons are radioactive p articles that act like clocks. They arrive in great numbers at the top of M ount Washington. (a)
They self-destruct at such a rate that very few would survive to reach sea level, if time in the muons’ reference frame were the
same as on Earth. (b) The muons exp erience time dilation, and thus most of them reach sea level. Detection of these many
muons confirms time dilation

Now, that 500 muons per hour figure isn’t unique to Mount Washington;
we would measure very nearly the same quantity each hour anywhere within
a few hundred miles if we were at the same 6,300-foot altitude as Mount
Washington. So here’s what the experimenters do: They go down to
Cambridge, Massachusetts, at sea level, and set up the same muon detector,
again looking for muons moving at 0.995c. They know that overhead, some
6,300 feet up, there are 500 such muons every hour coming into a region the
size of their detector. If they didn’t believe in relativity, then the
experimenters should expect to find very few muons at sea level, as shown in
Figure 9.2a.
Now let’s look at things from the muons’ point of view. They’re like the
clocks in our spaceship, which are present both as the ship passes Earth and
again when it reaches the star. The muons are present as they pass the 6,300-
foot altitude of Mount Washington, and—if they haven’t self-destructed—
they’re present again at sea level. So they’re clocks that are present at two
events, and for them the time between those events should be shortened
according to our time-dilation formula. For a speed of 0.995c, that formula
gives a relativistic factor equal to one-ninth. So the time to get from
Mount Washington to sea level, as judged by the muons, should be only one-
ninth what an observer at rest on Earth would expect. That muon time is so
much shorter that very few muons should decay on the way from 6,300 feet to
sea level.
We have a clear distinction between two possible outcomes. If relativity
is correct, and time dilation occurs, then very few muons will decay and the
experimenters should catch many muons in their detector at sea level (Figure
9.2b). If relativity isn’t correct, then most of the muons should decay on the
way down, and the sea-level experiment should find very few (Figure 9.2a).
So what happens? The result is clear: Nearly as many muons reach sea level
each hour as arrive at the top of Mount Washington. The muons are clearly
experiencing time dilation, which at their speed relative to Earth is a
dramatic effect. A look at the actual numbers in relation to the muons’
lifetime not only confirms that time dilation occurs but also that it obeys
precisely our time-dilation formula. Time for the muons elapses at one-ninth
the rate it does as measured by two clocks at rest on Earth, one on the
mountaintop and one at sea level.
We can also ask the same question here that we did for the spaceship.
How can the muons possibly get from Mount Washington’s altitude to sea
level in so short a time that they don’t self-destruct on the way? Doesn’t that
require them to be moving faster than light? No! Again, length contraction
comes to the rescue. Just as measures of time are different in the muons’
reference frame, so are measures of space. For the muons, the height of
Mount Washington is contracted by the same factor (one-ninth), so the 6,300-
foot mountain is, to the muons, only 700 feet high.
You may still be dissatisfied, because subatomic muons are so outside the
range of your everyday experience. But the muon experiment really is
completely analogous to our star trip, and we have every reason to believe
that what holds for the muons would hold for ordinary clocks, people, and
spaceships. If that weren’t the case, then we would have the Principle of
Relativity applying in the subatomic realm of physics but not to the physics of
ordinary objects. Think back to Chapter 2, and you should find that
possibility very unsettling.
Still, it would be nice if we could measure time dilation in ordinary
clocks big enough for us to see and hold. Well, we can—but the effect is less
dramatic because we can’t get regular-sized clocks to speeds that, relative to
us, are anywhere near the speed of light. Furthermore, because the effect of
time dilation is so small at the relative speeds we can achieve, we need
clocks that are good enough to mark extremely small time differences. We
have such clocks—they’re atomic clocks like those used to set the world
time standards. These clocks are accurate to a few billionths of a second. In
an experiment aimed at confirming relativistic effects on time, atomic clocks
that initially read the same time ended up disagreeing by several hundred
billionths of a second as a result of relative motion at the speeds of
commercial aircraft. We’ll revisit and elaborate on this experiment at the end
of the next section.

A Round-Trip Star Trip—The Twins Paradox

Making our star trip a round-trip brings us to one of the most famous results
of special relativity—the so-called twins paradox. Again we have the same
trip to a star 20 light-years distant, as measured in the Earth–star frame. Now
you board the spaceship and zoom off at 0.8c. Your twin sister remains
behind on Earth. When you reach the star you immediately turn around and
zoom back to Earth, again at 0.8c. You land, climb out of the ship, and greet
your sister. How do your ages compare?
We’ve already done the numbers for the outgoing trip: According to
clocks on Earth and star, the one-way trip to the star at 0.8c takes 25 years.
According to the ship clock, time dilation gives a one-way trip of only 15
years. What about the return trip? It’s identical to the outgoing trip. We still
have the situation described in Figure 9.1, except that the ship is going in the
opposite direction. Nothing in our careful analysis in Chapter 8 suggested
that the direction mattered. There we found that a clock that’s present at two
events reads less elapsed time between those events than do two
synchronized clocks in a frame of reference where the events occur at
different places. That’s as much the situation for the return trip as it was for
the outgoing trip. For the return, the two events are departure from the star
and arrival at Earth. They occur at the same place in the ship’s frame (right at
the ship, since you’re present at both events even though you never move
from your seat) and at different places in the Earth–star frame. So the time-
dilation formula applies just as much to the return trip, which we conclude
takes 15 years of ship time but 25 years of Earth–star time. Then the round-
trip takes 30 years of ship time but 50 years of Earth–star time. So, when you
return to greet your twin sister, you’re 20 years younger than she is!
This result is unambiguous. You’re standing right next to your sister and
all can see that she’s much older. There’s no question of someone observing
the two of you having to compensate for different light travel times or other
complications, because you’re right there beside each other. You really are
20 years younger than your sister, and no one can see it any other way.
By the way, things are a bit more ambiguous with the one-way trip,
because there’s the problem of communicating over the Earth–star distance to
establish your relative ages. If we send advance observers to the star, there’s
still the complication of time dilation affecting them—which is why I had
them sent well in advance in a much slower spaceship. We could, with care,
establish an unambiguous age difference if you travel to the star, stop, and
report your age. But it’s much clearer if we consider the round-trip, because
then you end up standing right next to someone who did nothing but stay at
rest relative to Earth while you were traveling.
Why is this result paradoxical? It’s still unsettling to think that you can
end up 20 years younger than your twin, but if you really accept the relativity
principle then you’re going to have to accept that age difference as well. The
paradox comes from considering the round-trip as viewed from the reference
frame of the spaceship. To you, onboard the ship, the trip starts with Earth
receding into the distance at 0.8c. Later, just as you’re at the star, Earth turns
around and comes back toward you. When the trip is finished you’re again
standing next to your twin on Earth. From your point of view it looks like you
stayed at rest in the ship, while Earth and your twin went off on a round-trip
at 0.8c. So why isn’t she the one who’s younger?
The resolution of this seeming paradox lies in what’s special about the
special theory of relativity. Remember that special relativity applies only to
the special case of reference frames in uniform motion. Remember even
further back to the philosophical shift from Aristotelian thinking to the ideas
of Galileo and Newton. For Aristotle motion was really absolute and to keep
something moving required a force. For Galileo and Newton, uniform motion
became a natural state of affairs, requiring no force or other cause. For
Galileo and Newton, it’s changes in motion that are important. Newton’s
laws quantify this new philosophical view, showing just how forces cause
changes in motion. Einstein’s special relativity builds on this Newtonian
idea, extending to all of physics the view that no one can make an absolute
claim to be moving or at rest. This is the essence of the special relativity
principle: that the laws of physics are the same no matter how you’re moving
—as long as you’re moving uniformly.
Now to the extent that Earth is in uniform motion—and to a very good
approximation, it is—your earthbound twin is in a reference frame in uniform
motion and the laws of physics work just fine for her. As long as you’re in
the spaceship moving uniformly, they work just fine for you, too. But when
the spaceship turns around at the star, and when it starts and stops at Earth, its
motion is decidedly not uniform. You, the ship’s passenger, know that
because you feel the force of your seatback accelerating you as the ship
leaves Earth, you experience the strong forces needed to turn you around at
the star, and you feel the ship braking when it returns to Earth. Your twin
back on Earth feels nothing as the ship accelerates, as it turns around, and as
it stops. So there really is a difference, in that the ship’s motion changes and
Earth’s doesn’t. It’s this difference that invalidates the argument from the
ship’s perspective. The ship is not in a single uniformly moving reference
frame during the entire round-trip, so we can’t expect the laws of physics to
apply, in the spaceship frame, to the entire journey. We can apply them to the
individual phases when the ship is in uniform motion, and that’s how we
conclude, correctly, that the ship time on each leg of the trip is only 15 years.
But we can’t say the ship is in uniform motion for the entire trip or that Earth
makes a round-trip including a turnaround when the ship is at the star. The
ship really does turn around, and Earth really doesn’t!
Do you sense an absoluteness about motion creeping in here? Yes, in that
I’m saying changes in motion are absolute in the context of special relativity.
Motion itself isn’t absolute but changes in motion are. Everyone agrees that
you feel strong forces when your spaceship turns around and that your twin
back on Earth doesn’t. The presence of those forces isn’t a relative thing;
they and the resulting change in the ship’s motion have an objective reality
that transcends one’s frame of reference. The difference between Earth and
ship perspectives on the round-trip is real, and it means that you and your
twin really are in different situations. Your argument that your twin sister
should be younger is simply invalid, because you’re applying the results of
special relativity in the nonuniformly moving reference frame of your
spaceship, where the physics of special relativity is not valid. So there’s no
paradox.
You might wonder why uniform motion is so special, and how we can
know for sure that a reference frame really is in uniform motion. Einstein
wondered too, and that helped lead him to his general theory of relativity—a
theory that removes special relativity’s restriction to uniform motion. We’ll
explore the general theory in the final three chapters and will revisit the
twins paradox in the context of general relativity.

Time Travel!
Our round-trip at 0.8c to that star 20 light-years away resulted in a 20-year
age difference between the traveling twin and the earthbound twin. Could we
travel in a way that would result in even greater age difference?
Suppose we can crank our spaceship up to a speed arbitrarily close to c.
Then the round-trip to the 20-light-year-distant star will take almost as short
a time as it would take for light—namely 40 years, round-trip. That’s the time
as measured in the Earth–star frame. But as the ship’s speed relative to Earth
and star approaches c, the speed v in the relativistic factor approaches
1 and the factor itself approaches 0. So the ship time, t' = t × ,
approaches 0 as well. If you’re in the ship, you can make that time an hour, a
minute, a second, or whatever you want, provided you set your speed close
enough to c. So the greatest age difference between you and your stay-at-
home twin would be just over 40 years. You get that maximum difference if
you go so close to c that (1) the time in the Earth–star frame is essentially the
same as the 40-year round-trip time for light itself and (2) the factor ,
and therefore the ship time, is very close to 0. In that case your star trip
amounts, for you on the spaceship, to a nearly instantaneous 40-year jump
into the future on Earth!
Forty years is the best you can do with a trip to that 20-light-year-distant
star. What if you go farther? For a star 50 light-years distant, the shortest
possible round-trip time, as measured by observers on Earth, is just over 100
years. Again, just how long it takes in ship time depends on exactly how
close your speed is to c. You can make that time 1 year, 1 week, 1 day, 1
second, or whatever you want. If you make it very short, then a round-trip to
the 50-light-year-distant star amounts to a jump 100 years into the future on
Earth. So if you’re really curious to know what things are going to be like on
Earth a century from now, you can find out, without getting significantly
older, by taking a round-trip at nearly the speed of light to a point 50 light-
years distant. Time travel into the future really is possible!
But there’s a catch. If you don’t like the Earth you find 100 years in the
future, there’s no going back. You’re either stuck where (or, rather, when) you
are, or you can take your chances on a jump further into the future. But you
can’t go back. That would cause all sorts of contradictions, as you know if
you’ve seen the film Back to the Future.
You can play this time-travel game with any numbers you like. The
farther you go in space, the farther you can jump into the future. A round-trip
to the center of our galaxy, 30,000 light-years distant, takes a minimum of just
over 60,000 years in the Earth’s frame of reference. But on a spaceship
making that trip, the round-trip could take a few years, a day, an hour, or even
less, depending on just how close v in t' = t × is to 1. So with a trip to
the galactic center you can jump 60,000 years into the future. If you travel at
nearly c to the Andromeda galaxy, 2 million light-years away, then you’ll
return to Earth 4 million years in the future.
Could it really happen? If you accept the Principle of Relativity, you have
to say yes. Experiments with muons and atomic clocks confirm time dilation,
the relativistic phenomenon at the basis of this futuristic time travel. Some
day real people, using real spacecraft, may travel seemingly impossible
distances in what are nevertheless minuscule times. If they choose to return to
Earth, they’ll have time traveled decades, centuries, millennia, or even
further into the future!

*The exp eriment is summarized by D. H. Frisch and J. H. Smith in American Journal of Phy sics 31 (1963): 242–355, and in
their film Time Dilation—An Exp eriment with mu-M esons (Educational Develop ment Center, 1963).
CHAPTER 10

THE SAME TIME ?

• • •

At this point you may still be a bit uncomfortable with relativity and its
strange consequences like different-aged twins and time travel to the future.
But I hope you see the logic that takes us directly from the Principle of
Relativity to these unusual implications. The relativity principle itself should
be firmly grounded in your mind; after all, it made complete sense in Chapter
2, and by Chapter 7 it had emerged from the quandaries of late-nineteenth-
century physics as the one clear way to resolve the muddle of motion and
ether.

A Test of Faith

Now I’m going to put your faith in relativity to the test. Consider once again a
one-way trip to that distant star. To make things really simple, forget about
starting at Earth and stopping at the star. Rather, the spaceship just comes
zooming past Earth at a steady 0.8c, then later zooms past the star. In other
words, the ship is always in uniform motion, so it’s always a good reference
frame in which to do physics.
The preceding chapter should have convinced you that, from the
viewpoint of observers on Earth, time on the spaceship “runs slow.”
Specifically, the time between the event of the ship passing Earth and the
later event of the ship passing the star is shorter in the ship’s frame of
reference than it is in the Earth–star frame.
 Now here’s my question: What do you, a passenger on the spaceship,
have to say about clocks in the Earth–star frame? To a friend on Earth, your
clock runs slow. Do you want to say, then, that from your shipbound
perspective, Earth and star clocks must run fast? Sure sounds like a logical
answer. After all, if your clock is running slow compared with clocks on
Earth and star, then surely Earth and star clocks must be running fast
compared with yours.
But wait! I’ve set things up in this example so that you in the spaceship
and your friend on Earth really are in equivalent situations, namely, the state
of uniform motion. Your friend says, “I’m at rest on Earth, and the spaceship
is moving relative to me, so the ship’s clock is running slow compared with
my clocks.” But your situation is perfectly equivalent, so you must be able to
say the analogous thing: “I’m at rest in the spaceship, and Earth is moving
relative to me, so Earth’s clocks are running slow compared with mine.”
Each observer—you in the spaceship and your friend on Earth—must claim
that the other’s clocks are running slow. How can that possibly be? Yet if you
really accept relativity—that the laws governing physical reality are the
same in all uniformly moving reference frames—then you’re forced to this
seemingly absurd conclusion.
You might still object that the spaceship is “really moving” while Earth
isn’t, so the situations aren’t really equivalent. But by now you’re surely past
such relativistically incorrect thinking! Unlike the round-trip we considered
in the preceding chapter, or even a one-way trip with starts and stops, the
scenario here really has both ship and Earth in completely equivalent
situations. Neither changes its motion in any way, so both are in the state of
unchanging, uniform motion that makes them both suitable reference frames
for applying the laws of physics. Or maybe you want to say that the ship is
moving in one direction relative to Earth, and Earth is moving in the opposite
direction relative to the ship, so that makes their situations different. But
nothing in our arguments leading to time dilation depended on the direction of
the relative motion. I could have redrawn Figure 8.3b in a reference frame in
which the light box was moving to the left, and the argument from that figure
would not have changed one bit. So the direction of relative motion can’t
matter.
So which is it? Do Earth’s clocks run slow or fast from the viewpoint of
the ship’s clock? The answer, if we accept relativity, has to be that they run
slow. To understand this answer, we have to find our way out of the
contradiction it seems to imply—the contradiction that, relative to Earth, the
ship clocks run slow while at the same time, relative to the ship, the Earth
and star clocks run slow. Escaping this contradiction will give you further
insights into the nature of time and especially the status of simultaneous
events, that is, pairs of events that occur at the same time.

Simultaneity Is Relative!
What does it mean to say that two events occur at the same time? If the two
events also occur at the same place, then there’s no question. If we see the
events simultaneously, then they must have occurred simultaneously. But
suppose the events occur at different places. Then an observer watching the
two events has to compensate for the time it takes light to get from each event
to determine if they were, in fact, simultaneous. You’ve already seen enough
of relativity to know that the constancy of the speed of light might make that
compensation different for observers in relative motion. So if one observer
determines that two events are simultaneous, it becomes an open question
whether another observer, moving relative to the first, will deem them
simultaneous.
Another approach to determining simultaneity is to set up clocks at the
locations of the two events. At each clock, station a reporter who will report
to you the time of the event occurring at the location of that clock. If both
times are the same, then you can claim that the events were simultaneous. For
this to work, you have to be sure those clocks are synchronized. What does
that mean? It means both clocks read the same time, say noon, at exactly the
same instant. In other words, the event of one clock’s hands pointing to noon
is simultaneous with the other clock’s hands pointing to noon. You could
synchronize the clocks by standing midway between them and sending a light
flash to your reporters at the clocks. Each reporter sets the clock at the instant
the light flash arrives. But again we’re back to judging simultaneous events
by methods that involve sending light signals through space and, again,
relativity leaves open the question of whether observers moving relative to
you will agree that the events are simultaneous. In this case, that means such
observers might not agree with you that your clocks are synchronized.
 In fact, events that are simultaneous in one frame of reference may not be
simultaneous in another reference frame. I’m now going to demonstrate this
rigorously in a way that follows from the Principle of Relativity. To do so,
I’ll invoke relativistic length contraction—a result which, as I showed in the
preceding chapter, follows logically from the relativity principle. Recall that
an object is longest in a reference frame in which it’s at rest and shorter as
measured in any other reference frame. In our earlier examples, one “object”
in question was the Earth–star pair, whose separation contracted from 20
light-years in the Earth–star frame to 12 light-years in the ship frame.
Another object we considered was Mount Washington, 6,300 feet high in the
Earth frame but contracted to only 700 feet in the reference frame of muons
for which Mount Washington was moving at 0.995c.
I’m now going to consider two distinct objects and look at how length
contraction applies to each of them in different reference frames. The objects
are two identical airplanes, and they’re flying toward each other at a
substantial fraction of the speed of light (these are no ordinary airplanes!).
The airplanes pass, one just above the other. I want to consider two events
associated with this passing. The first event will be the nose of the upper
airplane passing the tail of the lower one. We’ll call this event A. The second
event will be the tail of the upper airplane passing the nose of the lower one.
This is event B. We want to know whether events A and B are simultaneous.
Figure 10.1 shows the situation as observed in a reference frame in
which the two planes approach with the same speed. Obviously, neither
plane is at rest in this reference frame, so each is shorter, because of length
contraction, than it would be in a reference frame where it was at rest.
Because both are moving at the same speed relative to the reference frame of
Figure 10.1, each is contracted by the same amount. So in this reference
frame the two planes have the same length. When they’re alongside each
other, as in Figure 10.1b, the nose of the upper plane coincides with the tail
of the lower one at the same time that the tail of the upper plane coincides
with the nose of the lower one. That is, events A and B are simultaneous in
the reference frame of Figure 10.1.
Fig. 10.1 (a) Two identical airp lanes ap p roach. The figure is shown in a reference frame in which the p lanes ap p roach with the
same sp eed. Event A is the nose of the up p er p lane p assing the tail of the lower p lane; event B is the tail of the up p er p lane
p assing the nose of the lower one. (b) In this reference frame, events A and B are clearly simultaneous.

Now consider a reference frame in which the upper plane is at rest. In

this frame the upper plane has its greatest length. But the lower plane is
moving faster relative to the upper plane than it was relative to the reference
frame of Figure 10.1, so in the rest frame of the upper plane, the lower plane
is even shorter than it was in Figure 10.1. Figure 10.2a shows the situation at
the time the nose of the upper plane coincides with the tail of the lower plane
—that is, at the time of event A. Figure 10.2b shows the situation a little
while later. The upper plane hasn’t moved, since this is its rest frame. But
now the nose of the lower plane has reached the tail of the upper plane. This
is precisely our event B, so Figure 10.2b shows the situation at the time of
event B. The conclusion is obvious: In the rest frame of the upper plane, the
frame of Figure 10.2, event A occurs before event B.
Fig. 10.2 In a reference frame in which the up p er p lane is at rest, that p lane is longer than in Figure 10.1 because it isn’t length-
contracted. The lower p lane, however, is moving faster relative to this frame than it was in the frame of Figure 10.1, so it’s
even shorter. (a) The situation at the instant of event A; (b) event B. Clearly event A p recedes event B in this reference frame.

Finally, Figure 10.3 shows a reference frame in which the lower plane is
at rest. Now the lower plane is longer, the upper one shorter, and it’s obvious
that in this reference frame event B occurs before event A.
The example of the airplanes shows clearly that events that are
simultaneous in one reference frame may not be simultaneous in another
reference frame moving relative to the first. Comparison of Figures 10.2
and 10.3 shows something even harder to swallow: the order of two events
can be different in different reference frames. That might seem particularly
disturbing if you think about events that are causally related—like your birth
and your now reading this book. Are there really observers for whom the
reading comes before the birth? Fortunately, not. Causality is a big enough
issue that I’ll devote the entire next chapter to it; for now, let me just assure
you that not all pairs of events can have their time order reversed. All we’ve
learned from the airplane example is that those events that are simultaneous
in some reference frames may not be simultaneous in other frames, and it’s
precisely such events whose time order itself depends on the observer’s
frame of reference.

Fig. 10.3 Like Figure 10.2, but now in the reference frame in which the lower p lane is at rest. Here event B comes first.
Incidentally, Figures 10.1, 10.2, and 10.3 are scaled so that the sp eed of the two p lanes relative to the reference frame of Figure
10.1 is 0.6c.

By the way, I’ve been saying that events simultaneous in one frame “may
not be simultaneous” in another frame. Why “may not” as opposed to “are
not”? Are there cases where events can be simultaneous in different frames?
Yes, if the relative motion of the two frames is perpendicular to the line
joining the two events. For example, observers moving from the bottom
toward the top of the page in Figure 10.1 will judge events A and B to be
simultaneous, no matter how fast they’re moving relative to the frame of
Figure 10.1. For them, contraction of the airplanes squeezes them in the
vertical direction, not the horizontal, so it affects both planes the same way.

“Moving Clocks Run Slow,” Revisited

I now want to use your newfound knowledge about simultaneous events to

resolve the apparent contradiction in my assertion that earthbound observers
find that the spaceship clocks run slow, while observers on the ship find that
Earth’s clocks run slow. Look again at Figure 10.2, the airplane passings
viewed from the reference frame of the upper plane. In this frame the lower
plane’s tail passes the upper plane’s nose (event A) before the lower plane’s
nose passes the upper plane’s tail (event B). Here’s the point: in this
particular situation, in the reference frame of the upper plane, it’s the event
that takes place further to the right—that is, further in the direction from
which the “moving” plane is coming—that happens first.
Now, let’s apply this understanding to our star trip, where we found that
the trip time measured in the spaceship’s frame was shorter than the trip time
measured in the Earth–star frame. In the ship frame, we timed the trip with a
single clock in the spaceship. In the Earth–star frame, we timed it with two
different clocks, one located at Earth and the other at the star. Of course,
those clocks had to be synchronized, meaning that they both read 0 years at
the same instant. To be more precise, in the Earth–star frame the event of the
Earth clock reading 0 years is simultaneous with the event of the star clock
reading 0 years. Figure 10.4a shows all three clocks at the instant the ship
passes Earth, at which instant all three read 0 years. You can tell that the
figure is drawn from the viewpoint of the Earth–star frame, because it’s in
that frame that the events of the Earth clock reading 0 and the star clock
reading 0 are simultaneous. But we’ve just found that events simultaneous in
one reference frame aren’t simultaneous in another, so those events can’t be
simultaneous in the ship frame. What, then, is their time relationship?
The ship frame is like the frame of the upper airplane in Figure 10.2.
Ship and airplane are both at rest in their respective frames, and objects of
interest (the other plane, or the Earth and star) are moving toward them from
the right. Events that are simultaneous in one reference frame (the frame of
Figure 10.1 in the airplane example or the Earth–star frame in our star-trip
example) are not simultaneous in another frame (the frame of the upper plane
or the ship frame) and, furthermore, the event that is farther to the right occurs
first. That means the event of the star clock reading 0 occurs before the Earth
clock reads 0, as observed from the ship’s frame of reference. In other
words, the star clock is ahead of the Earth clock. So when the Earth clock
reads 0 time, the star clock reads a later time. That’s exactly what Figure
10.5a shows. Here, as observed in the ship frame, Earth (heading to the left)
passes the ship at the instant the ship and Earth clocks both read 0. But the
star clock is ahead, so it reads a later time. I won’t go through the math, but
that later time is, in fact, 16 years.
You can probably begin to see now how we’re going to get out of the
contradiction that each observer thinks the others’ clocks run slow. In the
ship frame the one-way trip takes 15 years. But I’m claiming that observers
on the ship must see Earth and star clocks running slow—and by the same
relativistic factor = 0.6 because the speed of Earth relative to the ship
is the same 0.8c as that of the ship relative to Earth. So to the ship’s
observers the clocks on Earth and star advance only (0.6) × (15 years) = 9
years. When the ship reaches the star (or, from the ship’s viewpoint, the star
reaches the ship), the star clock has advanced only 9 years. But, as observed
in the ship’s frame, the star clock already read 16 years when the ship passed
Earth. So now it reads 16 + 9 = 25 years—precisely the trip time as
measured in the Earth–star frame.

Fig. 10.4 Star trip of Figure 9.1, shown at two different times: (a) the situation at the instant the ship p asses Earth; (b) its
arrival at the star. Figure is drawn from the Earth–star frame of reference.
Fig. 10.5 Star trip , now drawn from the ship ’s reference frame. Here the ship is at rest, while the Earth and star move to the
left at 0.8c. Note that the ship is longer in its own frame, while Earth, star, and the distance between them are contracted. (a)
As the ship p asses Earth, both ship clock and Earth clock read 0. In the ship frame the Earth and star clocks aren’t
sy nchronized, and the star clock is, in fact, 16 y ears ahead. (b) The ship clock advances 15 y ears between the times Earth and
star p ass the ship , just as in Figure 10.4, but Earth and star clocks are “running slow,” and they advance only 9 y ears. Because
it was 16 y ears ahead, the star clock still reads 25 y ears as the star p asses the ship , just as in Figure 10.4.

Let’s sort all this out. Turn back to Figure 10.4, where (a) shows the
situation in the Earth–star frame when the ship passes Earth and (b) shows
the situation in this frame when the ship passes the star. In this frame all the
clocks read 0 as the ship passes Earth (Figure 10.4a). Later, as the ship
passes the star, its clock reads 15 years while Earth and star clocks—which
are synchronized in the Earth–star frame—both read 25 years (Figure 10.4b).
The interpretation of the clock readings in the Earth–star frame is that 25
years have elapsed in that frame but only 15 years have elapsed in the ship
frame because the ship clock runs slow.
Now look again at Figure 10.5, where (a) shows the situation in the ship
frame when Earth passes the ship and (b) shows the situation in this frame
when the star passes the ship. As in the Earth–star frame, both ship clock and
Earth clock read 0 as Earth and ship pass. But in the ship frame, the clocks in
the Earth–star frame aren’t synchronized; in fact, the star clock already reads
16 years when Earth and ship pass. Later, as the star and ship pass, the ship
clock has advanced from 0 to 15 years. The star clock, running slow from the
viewpoint of the ship frame, has advanced only 9 years. But since it was
ahead to begin with, it now reads 25 years. The interpretation of the clock
readings in the ship frame is that 15 years have elapsed in that frame while
only 9 years have elapsed in the Earth–star frame.
Yet observers in both frames agree about what the clocks actually read
whenever two clocks are right next to each other so observers can
unambiguously compare times. In particular, observers in both frames agree
that 15 years elapse on the ship’s one clock. They also agree that the reading
of the Earth clock at the Earth/ship passing differs by 25 years from the
reading of the star clock at the star/ship passing. What they disagree about is
the interpretation of this 25-year interval. To observers in the Earth–star
frame the clocks at Earth and star are synchronized, so 25 years is a
legitimately measured time between two events. To observers on the ship,
Earth–star clocks are running slow, and advance only 9 years between the
two events. The 25-year difference in clock readings occurs because the star
clock is ahead of the Earth clock by 16 years.
Which interpretation is right? By now you should have enough faith in
relativity to know that’s a meaningless question. Observers in both frames
are correct in asserting that the other’s clocks run slow, and they’re saved
from outright contradiction by the fact that events simultaneous in one
reference frame aren’t simultaneous in another frame in motion relative to the
first.
It’s easy to conjure up situations in relativity that seem contradictory, and
books on relativity are often full of paradoxes that seem to embody such
contradictions. But there’s always a way out, and it usually involves
recognizing that simultaneity is relative. Careful consideration of the timing
of events in two different reference frames almost always resolves any
apparent contradiction.
CHAPTER 11

PAST, PRESENT, FUTURE, AND . . . ELSEWHERE

• • •

That simultaneity is relative has just got us out of the seeming contradiction
of observers in relative motion each finding that the other’s clocks “run
slow.” But in demonstrating the relativity of simultaneity, I introduced what
may seem an even more disturbing thought—that the time order of events may
depend on one’s frame of reference. Because relativity gives every uniformly
moving reference frame equal status, this reversal of time order isn’t just
some illusion. It’s really true that I can observe event A to occur before B,
that you can observe B before A, and that we’re both right. But how can that
be? Doesn’t it wreak havoc with causality?

The Past Is History

Ask a historian what the past is and you’ll get an answer something like “all
those events that have already occurred.” What does it mean that an event has
already occurred? It means that the time of that event is earlier than the time
of the present event. But we’ve just seen that the time order of events may
depend on one’s reference frame, so how can the notion of the past have any
meaning at all? Or for that matter, the future?
There are, in fact, events that are unambiguously in the past—meaning
that they occurred at a time before the present event. Note that I didn’t say
“before the present.” I said “before the present event.” That’s because the
relativity of simultaneity precludes my talking about a universal present
instant that pervades the whole Universe. Back to the star-trip example: For
an observer on Earth at the instant the ship passes, the present includes the
events of Earth’s clock reading 0 and the star’s clock reading 0. But to an
observer in the spaceship at the same instant, Earth’s clock reading 0 is
indeed an event in the present, but the star’s clock reading 0 is not—since
that event occurred 16 minutes earlier! So there’s no such thing as a universal
“present.” There is, for me, the present event—namely, whatever is occurring
here and now. “Now” isn’t enough; I have to indicate “here” as well—and
that means I’m talking about an event, not just a time.
What are some events that are truly in the past, meaning they
unambiguously occurred before your present event, that is, the event of your
reading these words? For one, your birth. There are no observers, in any
state of motion, who would judge that event to occur after your here and now
(although different observers will disagree about the amount of time between
those events). We don’t have to restrict ourselves to events in relation to the
here and now. We can also ask, for example, whether the event of the Titanic
hitting the iceberg preceded the event of the great ship’s sinking. The answer
is an unambiguous yes. Again, one event is clearly in the other’s past.
Consider also that in 1987 astronomers observed a supernova—an exploding
star—in a neighbor galaxy some 160,000 light-years away. Clearly the
supernova event itself occurred before the astronomers observed it, since it
took light from the supernova 160,000 years to reach the astronomers’
telescopes.
What do the three pairs of events we’ve just considered have in
common? They’re all causally related. Your birth is a necessary cause of
your reading these words. Had the first event not occurred, the second could
not have occurred either. Had the Titanic not hit the iceberg, it would not
have sunk. Had the supernova explosion not occurred, the astronomers would
not have observed it. In each case, the earlier event was capable of
influencing the later one and, in fact, did influence it. That provides a more
robust definition of the past: The past of a given event consists of all those
events that are capable of influencing the given event. Similarly, the future of
the given event consists of all those events that the given event can influence.
Note again that I’m talking about past and future in relation to a specific
event; in a Universe in which simultaneity is relative, there’s simply no such
thing as a universal past and a universal future. But when one event is in
another’s past, that relationship is not ambiguous. All observers will agree
about which event came first (although, again, they may disagree on the
amount of time between the events). So relativity doesn’t violate causality, in
that those events that are causally related have an absolute time order that no
observer will dispute.
But are all pairs of events causally related? To get at that, I’ll ask some
questions that seem, at first, very strange: Are there events that have already
occurred that are not in the past? Are there events that haven’t occurred yet
that are nevertheless not in the future? With our traditional definitions of past
and future, the answer to both questions is obviously no, but with our “can
influence” definition, the answer is less obvious. Are there events that have
already occurred but that cannot have any influence on my here and now? Are
there events that have not yet occurred but that cannot be influenced by what
I’m doing right here and now?

Excitement on Mars

When NASA’s Mars Rover was exploring the Martian surface in 1997, Mars
was 10 light-minutes from Earth. That means it took light, as well as the
radio waves used to communicate with Rover, 10 minutes to get from Earth
to Mars or from Mars to Earth. You’ve probably heard that the speed of light
is the maximum possible speed in the Universe—a consequence of relativity
that I’ll explore in detail in the next chapter. For now, let’s accept this cosmic
speed limit which, more accurately, states that information cannot be
communicated at speeds faster than c. So that 10-minute travel time for light
between Earth and Mars is the shortest time that any information could take
to travel between the two planets.
Now, let’s put ourselves in NASA’s Mars Rover control room, during the
time of the Rover mission. To make things concrete, let’s say we’re just
beginning a coffee break. Suppose further that, 5 minutes before our break
began, Rover’s TV camera spotted a real, live Martian strolling across its
field of view. Is that event in our past? It happened 5 minutes ago, so it
occurred before our present event. But could it influence us right now? Can
we possibly know, right now, of Rover’s discovery? No, and we won’t be
able to know for another 5 minutes, when we receive Rover’s TV picture. So
Rover’s discovery of extraterrestrial life cannot possibly influence us, right
now at the start of our coffee break here in NASA’s control room. We, the
Rover scientists, aren’t excited by what’s happened, because we simply can’t
know about it. By our “influence” definition, then, Rover’s discovery is not
in our past. It cannot be a cause of the event, namely our coffee break, that’s
occurring right now in the earthbound Rover control room.
We have in Rover’s discovery on Mars and our earthly coffee break 5
minutes later a pair of events that just cannot be causally related, because no
information can travel between them. They’re too far apart in space and too
close in time for even light to be present at both events. It’s precisely such
pairs of causally unrelated events that different observers, moving in this
case relative to our Solar System, may see with different time orders. For
example, there could be an observer for whom the two events are
simultaneous and others for whom the coffee break occurs first. Still other
observers would agree with us on Earth that the Martian discovery occurred
first, but they wouldn’t agree that the time interval was 5 minutes. I won’t go
into the mathematics of all this. But imagine in Figure 10.5 replacing Earth
with Mars and the star with Earth. Then you can see that an observer moving
in the direction from Mars toward Earth will observe Earth time advanced
relative to Mars time. That advancement will be 5 minutes—meaning the two
events will be simultaneous—if the observer happens to be moving at half
the speed of light. Observers moving faster in the same direction will judge
the coffee break on Earth to occur before Rover’s discovery of Martian life.
Observers moving slower but in the same direction will judge the Mars event
to occur first, but by less than 5 minutes.
Of course all these observers are correct, because they’re all in
uniformly moving reference frames that are equally valid situations for
applying the laws of physics. As you well know by now, measures of time
simply aren’t absolute, but depend on one’s reference frame. What’s new
here is that the time order of events may also depend on reference frame. But
there’s no contradiction. That’s because the only events for which different
observers claim different time orders are those that can’t be causally related,
because they’re too far apart in space and too close in time for any influence
traveling at c or less to get between them. Too far apart by whose measures?
By anyone’s. Even though the different observers disagree about the time
between the events, and for that matter about the distance between Earth and
Mars, all agree that the distance is greater than the distance light could travel
in the time between the events. More on this in Chapter 13.
Poor Mars Rover! It’s made a great discovery, but, alas, its navigation
system has failed and it’s heading straight toward a deep crater. We, in the
NASA control room, realize this just as we start our coffee break, and we
calculate that in 5 minutes Rover will topple into the crater and be destroyed.
Is Rover’s demise in our future? Can we send a radio signal to stop Rover
and prevent the impending disaster? No; Rover is 10 light-minutes away, and
we have only 5 minutes until Rover reaches the crater. Nothing we do at our
present moment can influence what happens on Mars 5 minutes from now. So,
according to our influence definition of the future, the event of Rover falling
into the crater is not in our future. Here’s another pair of events—the start of
our coffee break and Rover’s unfortunate accident—that cannot be causally
related. And, sure enough, there could be other observers, moving relative to
Earth and Mars, who find these events simultaneous, and still others for
whom Rover’s demise occurs before our coffee break starts.

The Elsewhere
If the events of Rover’s alien discovery and its crater accident aren’t in our
past or our future, where are they? Relativity opens a new realm of time—or,
more precisely, of spacetime. Events that aren’t in the past or the future of a
given event are in its elsewhere. They’re events that cannot communicate
with the given event, so the two cannot be causally related—and different
observers can disagree about their time order. On Mars, for example, any
events occurring between 10 minutes before and 10 minutes after the start of
our NASA coffee break are in the elsewhere of the coffee-break event.
Similarly, events in the 2-million-light-year distant Andromeda galaxy are in
our elsewhere if they occurred anytime more recently than 2 million years
ago or will occur within the next 2 million years. Even events that occur on
the other side of the room you’re sitting in are in your elsewhere, provided
they occurred less than a few nanoseconds (billionths of a second) ago or
will occur less than a few nanoseconds from now. Clearly, the farther away a
place is, the broader the range of time that lies in the elsewhere of your here
and now.
 You can picture the elsewhere with the help of Figure 11.1. Here I
assume for simplicity that we live in a space with just one dimension—
meaning we can move back and forth along a line but not in any other
direction (it gets too hard or even impossible to draw with more
dimensions!). I’ll represent place—that is, where an event occurs—by its
position along a horizontal line. I’ll represent time—when an event occurs—
by its position along a vertical line. To make things simple, I’ll measure
position in light-years and time in years. Every point on the diagram
represents a specific place and a specific time so the diagram is therefore a
spacetime diagram. But time and place are what specifies an event, so each
point on the spacetime diagram marks an event. The point at the center, where
the time and space axes cross, is the here (position 0) and now (time 0).

Fig. 11.1 Sp acetime diagrams, showing only one sp atial dimension. Each p oint in the diagram rep resents an event, its location
in sp ace given by its horizontal p osition, time by its vertical p osition. (a) In the Newtonian view, the p resent is the horizontal
line sep arating future from p ast. All events hap p ening right now, no matter where they are, occur in the p resent. (b) In
relativity, future and p ast have different meanings for each event. Diagram shows the p ast, future, and elsewhere for event A.
Tick marks are at intervals of 1 light-y ear in sp ace and 1 y ear in time. The p ast comp rises all those events that can influence A;
the future those that A can influence. The text discusses the relations among the six events A–F.

So far, there’s no relativity in all this, just a way of locating events on a

simple diagram. Before relativity, people believed in a universal present that
was the same instant for all, no matter where or what their state of motion.
Figure 11.1a shows that situation. Any event at the present time lies on the
horizontal line marked “present.” Events below this line are at earlier times,
or, according to our prerelativity definition, in the past. Events above the line
occur at later times and are therefore in the future.
But relativity precludes information transmission faster than the speed of
light and, therefore, puts a whole class of events into the elsewhere of the
present event. Figure 11.1b shows the relativistic situation. I’ve marked six
events on the diagram. To help locate them, I’ve put tick marks at intervals of
1 light-year on the horizontal (space) axis and 1 year on the vertical (time)
axis. Event A is the here and now, or the present event (time 0, place 0).
Event B occurs 2 years from now, and 1 light-year to the right of the present
event. Since it’s only 1 light-year away in space, there’s plenty of time for
light (or something slower, like a spaceship moving at half the speed of light)
to communicate information from event A to event B. That is, A can influence
B, and so B is unambiguously in A’s future. What about event C? It occurs 3
light-years to the right of A, but only 1 year away in time. There’s no chance
for light or anything else to get from event A to event C. Thus A cannot
influence C, so C is not in A’s future but in its elsewhere.
Figure 11.1b also shows a pair of 45-degree lines through point A. Any
event above A and between these lines is closer to A in distance (as
measured in light-years) than it is in time (as measured in years); therefore A
can influence such an event. All the events above A that are within the cone
defined by these lines are in A’s future. Similarly, an event like D that
occurred 2 years before A and 1 light-year away is clearly in A’s past, along
with all other events lying below A and within the cone defined by the 45-
degree lines. Finally, events like C, E, and F that occur farther to the right or
left than they do above or below A are in A’s elsewhere. That C occurs after
A, E before A, and F simultaneously with A is of no consequence because
these pairs of events can’t be causally related.
I’ve drawn Figure 11.1b from the viewpoint of a particular reference
frame. Observers moving relative to this frame would have different
measures of space and time, and their diagrams would be different. All
observers would have events B and D in the future and past, respectively, of
A, although how far away they are in time and in space would be different.
All observers would have C, E, and F in A’s elsewhere, but some would
have C below the horizontal axis and some would have E above it. That is,
for some observers event C would have occurred before A, and for some it
wouldn’t have. Still others would have A and C occurring simultaneously;
others, A and E. And for some, F would occur after A; for others, before. But
all that’s OK because C, E, and F are in A’s elsewhere and can’t be causally
related to A.
Note that I’ve been careful to say “the future of event A” or “in event A’s
elsewhere” but not “in the future” or “in the elsewhere.” That’s because the
terms “past,” “future,” and “elsewhere” are meaningful not in relation to a
time alone, but in relation to an event. In Figure 11.1b, for example, event B
is in event A’s future, but it’s not in event C’s future because it’s 2 light-years
to the left but only 1 light-year later in time. Drawing diagonal lines through
C rather than through A would make this clear. Similarly, event E is in A’s
elsewhere but in D’s future, while F is clearly in B’s past.
You might wonder why the concept of the elsewhere isn’t more obvious,
given the large region it occupies in Figure 11.1b. Again, the answer lies in
our limited experience. We don’t tend to move around at high speeds relative
to our surroundings or to deal with objects so distant that the travel time for
light is significant. And we measure distances in, say, miles, and times in
hours. In those units the diagonal lines in Figure 11.1b would not be at 45
degrees, but at a very low angle, hardly distinguishable from the horizontal
axis. That’s because those lines represent the speed of light, which is a great
many miles of distance for each hour of time. So the picture in our everyday
units of miles and hours would look much more like the Newtonian view of
Figure 11.1a. Only at great distances to the right or left of the here and now
point would we notice that the very slightly sloping speed-of-light lines left a
small wedge of elsewhere between past and future.
The elsewhere may sound like some mysterious new realm that’s forever
inaccessible to you. But it’s not; for you, the elsewhere comprises those
events that can’t influence or be influenced by you as you are here and now.
An event 5 minutes ago on Mars is in your elsewhere if you’re here on Earth,
but 5 minutes from now it will be in your past and at that time you can know
of the Martian event and be influenced by it. Similarly, you can’t do anything
to influence an event that’s going to occur 5 minutes from now on Mars, so
it’s in the elsewhere of your here and now. But if you had known an hour ago
(or any time more than 5 minutes ago), you could have taken steps to prevent
or alter the Martian event. That is, it was at one time in your future. But now
it’s in your elsewhere, and 15 minutes from now it will be in your past. Its
relation to you changes because you’re not an event. Its relation to a fixed
instant in your life doesn’t change, because that relation is truly between two
events.
If you’re not an event, then what are you? On a spacetime diagram, you’re
a path, called a worldline. Each moment in your life occurs at some time and
some place, so each moment represents an event, or a point in a spacetime
diagram. Since you can’t jump abruptly to a different time or a different
place, the events in your life form a continuous sequence of spacetime points.
The sequence advances inexorably into the future, but you’re free to move
about in space (in reality, you can move in all three dimensions, but only left
or right in my simplified spacetime diagram). Figure 11.2 shows a spacetime
diagram with the worldline of a human life. Note that no part of the worldline
path can be inclined at 45 degrees or more to the vertical in a diagram scaled
like Figures 11.1b and 11.2, since that would imply travel at light speed or
faster. Put differently, each point on an object’s worldline must be in the
future of the points preceding it. What about a light beam? Its worldline is a
45-degree line, showing that light advances 1 light-year in space for each
year of time. In more advanced treatments of relativity, worldlines play a
crucial role in describing occurrences ranging from our star-trip example to
the plunge into a black hole.
Fig. 11.2 The worldline of a human life is the continuous sequence of events between birth and death. This individual was born
some distance to the left of where she is now and is just under halfway through her life. Can y ou find a p eriod when the
individual stay ed p retty much in the same p lace? Note that the worldline is never at more than 45 degrees to the vertical,
meaning that all events in the individual’s life are causally related to the life events that p recede them.
CHAPTER 12

FASTER THAN LIGHT ?

• • •

The premise behind the past-present-future-elsewhere structure of spacetime

that I introduced in the previous chapter is that information cannot be
transmitted faster than the speed of light, c. Were that premise wrong—in
particular, if the instantaneous transmission of information were possible—
then past and future would be unambiguous and there would be no elsewhere.
Furthermore, with instantaneous transmission of information, we could
synchronize all clocks everywhere with a single instantaneous signal, and
there would be a universal time for everyone after all. Finally, events that we
thought could not be causally related actually could be. Yet relativity tells us
that those events can have their time orders reversed, which means that for
some observers effect would precede cause! You can see that faster-than-
light communication would wreak havoc with causality. Fortunately, it’s not
possible. Why not?

First Principles
First I’m going to give a quick and easy answer, one that follows directly
from the Principle of Relativity. It won’t be particularly satisfying, but it will
be correct. In the sense that all the implications of relativity follow from the
principle, it will be as complete as can be. Later, though, I’ll provide
answers you may find more satisfying.
The relativity principle states that the laws of physics are the same in any
uniformly moving frame of reference. Among those laws are Maxwell’s
equations of electromagnetism, and among the predictions of Maxwell’s
equations is the existence of light waves going at speed c. That prediction
must hold in all reference frames, meaning that all observers must measure
the same speed c for light. I’ve used that argument many times before, so it
should be tediously familiar.
Now suppose you and I are standing together as a light beam goes by.
You hop into a fast rocket and try to catch up with the light. If you succeed,
then you’ll be moving with speed c relative to me, and you’ll be at rest with
respect to the light. If you were at rest with respect to the light, then you
would be measuring a speed for light, namely zero, which is not equal to c.
But light is supposed to have speed c with respect to any uniformly moving
reference frame. If it didn’t, the Principle of Relativity would be violated. So
your situation—moving with speed c relative to me and therefore being at
rest with respect to the light—must be impossible.
This question of catching up with light puzzled Einstein from the age of
16 until he resolved it with his special theory of relativity. He imagined
running alongside a light wave, so the wave would be at rest with respect to
him. Einstein understood Maxwell’s electromagnetism, and he knew that
Maxwell’s equations dictated a particular form for an electromagnetic wave,
a form in which electric and magnetic fields are perpendicular to each other
and to the direction of the wave’s motion. What bothered Einstein was that a
stationary structure of that form was not a valid solution of Maxwell’s
equations. In fact, the only valid solution had the electromagnetic wave
moving at speed c. So there had to be something wrong with the idea of
running alongside a light wave so that the wave would appear to stand still.
Before relativity, there was no real problem here. Maxwell’s equations
were thought to be valid in only one particular frame of reference, namely the
frame at rest with respect to the ether. You wouldn’t expect electromagnetic
waves as seen from other frames to satisfy Maxwell’s equations, and there
would be no problem with an observer moving through the ether in such a
way that an electromagnetic wave appeared at rest.
But take away the ether, as Einstein did, and require that the laws of
physics be valid for all uniformly moving observers. In other words, insist
on the Principle of Relativity. Then all observers must measure c for the
speed of electromagnetic waves, and no observer can be at rest relative to
such a wave. Thus the Principle of Relativity asserts that it’s impossible for
you to run alongside a light wave and see it at rest, so it’s impossible for you
to move relative to me at speed c.
This argument is logically airtight and follows directly from the relativity
principle. But it isn’t very satisfying, because it doesn’t explain why you
can’t get yourself up to the speed of light. What actually prevents that? I’ll
now explore several more satisfying answers to that question. However,
there’s no new physical principle involved; like everything else in relativity,
these answers, too, are ultimately grounded in the Principle of Relativity.

Leapfrogging to c
Here’s an obvious way to beat the cosmic speed limit, c. Suppose we have a
big rocket capable of going, relative to Earth, at three-quarters of the speed
of light (0.75c). Inside the big rocket, we build a miniature version with the
same technology. We fire up the big rocket and zoom away from Earth at
0.75c. You climb into the small rocket, fire the engine, and soon you’re
moving at 0.75c relative to the big rocket (Figure 12.1). So now you must be
moving at 1.5c relative to Earth, right?

Fig. 12.1 The big rocket is moving p ast Earth at 0.75c. A smaller rocket moves at 0.75c relative to the big one. So is the smaller
rocket going at 1.5c relative to Earth? Common sense suggests it is, but in fact the sp eed of the small rocket relative to Earth is
only 0.96c.
 Wrong! Why wrong? Because measures of space and time aren’t the same
in different frames of reference. The small rocket is indeed moving at three-
quarters of light speed relative to the big rocket, and the big rocket is moving
at three-quarters of light speed relative to Earth. But determining the speed of
the small rocket relative to the big one involves measurements of distance
and time in the reference frame of the big rocket. An observer on Earth
doesn’t agree with those measures and comes up with a different result for
the speed of the small rocket. I won’t go through all the math, but that speed
is, in fact, 0.96c.
What I’m saying here is that 0.75c and 0.75c don’t add to give 1.5c.
That’s an odd statement! Surely if I jog at 5 miles an hour down the aisle of
an airplane going 600 miles per hour relative to the ground, then I’m going at
605 miles an hour relative to the ground. Why isn’t it the same for the
rockets? Actually, it is the same in both cases, but it’s the example of the
airplane that’s wrong. In fact, my speed relative to Earth is a tiny bit less than
605 miles per hour—and for the same reason, namely that measures of time
and space in the airplane’s reference frame aren’t quite the same as on Earth.
Here the difference is negligible, because the relative speed of Earth and
airplane is tiny compared with the speed of light. The effect becomes more
dramatic as relative speed increases.
To get a little more abstract, suppose that the big rocket is moving at
some speed u relative to Earth, and the small rocket is moving at speed v
relative to the big one. Common sense suggests that the speed of the small
rocket relative to Earth—call it v’—should be just v’ = u + v. But relativity
modifies this, giving

For obvious reasons, this equation is called the relativistic velocity addition
formula. Here I’m assuming that all speeds are given as fractions of the
speed of light. The numerator in the formula is just what we’d expect from
common sense; it’s the sum of the two speeds u and v. But the denominator
has the effect of reducing that value. That effect is small if either of the
speeds u or v is much less than the speed of light, since then the product u × v
is much less than 1 and the denominator remains essentially 1. At high
relative speeds, though, u × v is substantial and the speed of the small ship
relative to Earth is a lot less than common sense would suggest. For example,
let both u and v be 0.75 as in our example, and you’ll see that the result is v’
= 0.96.
It won’t help to make the small rocket go even faster. No matter what its
speed relative to the big rocket, as long as that speed is less than c then its
speed relative to Earth will also be less than c. We can also ask what will
happen if you shine a light beam inside the big rocket. Obviously, the light
goes at c relative to the big rocket. What’s its speed relative to Earth?
Relativistic velocity addition tells us that, too. The light takes the place of the
small rocket, so we put 1 in for v in our relativistic velocity addition formula
(that’s 1 because we’re measuring all speeds as fractions of c, so the speed
of light itself is just 1):

In other words, the speed of the light is unchanged. It’s the same relative to
Earth as it is relative to the rocket, namely 100 percent of c. Of course, we
already knew this; invariance of the speed of light is at the basis of relativity.
Relativistic velocity addition, like all else, is based on the Principle of
Relativity, so of course it gives a result consistent with that principle.
So we can’t leapfrog to a speed greater than c by adding two sub-c
motions. That rules out another idea you might have for achieving faster-than-
light travel. Think of taking one of those people movers found in airports and
put another people mover on top of it. People on the second mover should be
going twice as fast relative to the ground as those on the first. Of course, that
still isn’t very fast. So pile more and more people movers, one on top of the
other. Get enough of them and folks on the uppermost one should be going
faster than c relative to the ground. You should be able to get to that
superluminal speed by jumping from one mover to the next, undergoing just a
tiny increase in speed each time. But it won’t work. Even on the second
mover, you aren’t going quite twice as fast as people on the first one. That
slight discrepancy compounds, according to relativistic velocity addition, so
no matter how many people movers you pile up, none will be going faster
than c relative to the ground.
 Incidentally, understanding relativistic velocity addition helps dispel the
common misconception that it’s only light that exhibits unusual behavior. The
unusual behavior of light is, of course, that it has the same speed for all
observers regardless of their states of motion. For an object moving at less
than c relative to one observer, another observer in motion relative to the
first will measure a different speed for the object. This makes it seem that the
unusual behavior—invariant speed—is unique to light. But as the relative
speeds of observers and object approach c, the speeds measured by the two
observers become nearly the same and their difference becomes much less
than the relative speed of the two observers. So this relativistic velocity
addition effect applies to everything, not just light. The only distinction is that
with light there’s no difference whatsoever in the speed as measured by
different observers. For objects with relative speeds below c, there is a
difference, but that difference becomes very small as the speed approaches c.
Again, the effect here is not about light but about something more
fundamental, namely, the nature of space and time.

E = mc2

E = mc2 is surely the most famous equation in all of physics, and for most
people it’s synonymous with Einstein and relativity. It’s also, in the popular
mind, the basis of nuclear weapons. So what’s E = mc2 doing here, a seeming
footnote buried deep in a chapter on why faster-than-light travel is
impossible?
In fact, E = mc2 is itself a footnote to special relativity. It doesn’t appear
in Einstein’s famous paper of June 1905. Later that year, Einstein published a
second paper in which he asserted, “The mass of a body is a measure of its
energy content,” or, stated mathematically, E = mc2. He elaborated more fully
on this relation between mass and energy in a 1907 paper.
E = mc2 is important but it’s not the essence of relativity, and it’s no more
the basis of nuclear weapons than it is of a burning candle or a cave
dweller’s fire. Here I’m going to explore the meaning of E = mc2, then come
round to show how it gives what is probably the most satisfying reason that
faster-than-light travel is impossible.
 E = mc2 asserts a fundamental equivalence, or, more precisely, an
interchangeability, between matter and energy. Before Einstein, matter and
energy were the two substances that populated the physical Universe. I use
the word “substance” here in the sense of being “substantial”—that is, having
ongoing, indestructible, existence. Matter is obviously substantial; it’s the
stuff we and everything else seem to be made of. Matter is quantified by its
mass, a measure, roughly, of how much of it there is. We can transform matter
from one form to another, as in chemical reactions. Burning coal is an
example; carbon in the coal combines with oxygen in the air to make carbon
dioxide. The total amount of matter seems unchanged; what’s happened is a
rearrangement of atoms to make a new chemical substance. But the atoms
themselves seem unchanged. In other words, matter seems to be conserved; it
can be rearranged, but not created nor destroyed.
Energy is less tangible than matter, but it too was thought in prerelativity
times to be conserved. Like matter, energy can change forms. Sunlight falls
on a dark surface, transforming what was the energy of electromagnetic
waves (light) into the random molecular motions we call, loosely, heat.
Falling water turns an electric generator, transforming the energy of motion
into electrical energy. Coal burns, transforming chemical energy into heat.
You charge a battery, turning electrical energy into stored chemical energy.
You put the battery in your laptop computer, and the chemical energy changes
back into electrical energy to run the computer. You step on your car’s
brakes, changing the energy of the car’s motion into useless heat energy
(unless you’ve got a hybrid or electric car, in which case the energy goes
back into the battery for later use). I could go on and on, since
transformations among different forms of energy are involved in virtually
everything that happens. In all these transformations, it seems that the total
amount of energy remains unchanged.
What E = mc2 says is that matter and energy are interchangeable. A piece
of matter with mass m could, in principle, be transformed into pure energy,
where the amount of energy, E, is the product of the mass m and the square of
the speed of light. Because c is so big, this means that a little bit of matter
could yield a large amount of energy. The equation works the other way, too.
It says that an amount of energy E could be transformed into matter with mass
m given by E/c2. So matter and energy are, individually, no longer conserved.
What is conserved is a new universal substance, which for lack of a better
name we might as well call mass-energy. The total amount of mass-energy
remains the same, but how much of it is in the form of mass and how much of
it is energy can change.
Is the conversion of matter to energy and vice versa really possible? Yes,
but total conversion occurs only in very special cases. It turns out that every
one of the elementary particles that makes up everyday matter has an
associated antiparticle, identical in mass but opposite in electric charge and
other properties. Corresponding to the negatively charged electron, there’s a
positively charged antielectron, also called a positron. Corresponding to the
positively charged proton is a negative antiproton. Collectively, these
antiparticles constitute antimatter. Our Universe, for reasons that are not yet
entirely clear, seems to consist almost entirely of matter. A few antimatter
particles are created in high-energy collisions and in some nuclear reactions,
either naturally or in laboratory experiments. When a matter particle and its
antimatter opposite meet, the two annihilate, disappearing altogether in a
burst of energy. If the particle and antiparticle each have mass m, then the
total mass that disappears is 2m, and the energy that appears in its place is
therefore 2mc2. The opposite process can occur, too. Under the right
conditions, energy, in the form of a pulse of electromagnetic waves, can
disappear and in its place a particle and its antiparticle come into existence.
Figure 12.2 shows such a pair-creation event, as observed with a detector
used in high-energy physics experiments.
Complete conversion of matter to energy is such a powerful process that
if you had a box of ordinary raisins and a box of antiraisins (much harder to
get hold of!), the energy released in the annihilation of one raisin–antiraisin
pair would be enough to supply all of New York City’s energy needs for a
day. If you had a power plant capable of harnessing that energy, you could
drop in one raisin and one antiraisin each day and that would be all the fuel
you’d need to keep New York going.
Fig. 12.2 Simulated image of a p air-creation event, as observed in the p article detector of a high-energy p hy sics exp eriment.
The detector records the p aths of electrically charged p articles. An uncharged, and therefore invisible, gamma ray –a high-energy
bundle of electromagnetic energy –has entered the region from below. At the p oint marked, it ceases to exist and its energy
becomes that of an electron and its antip article,a p ositron. A magnetic field p oints p erp endicular to the p age and causes the
two new p articles to sp iral in op p osite directions.

Complete conversion of matter to energy is a rare event in the present

Universe. (However, in the earliest instants of the Big Bang both it and the
reverse were common and important processes.) But E = mc2 isn’t just about
matter–antimatter annihilation and other exotic happenings. Anytime a
process results in release of energy, there’s a corresponding reduction in
mass. Weigh a candle, for example, and then light it. Weigh all the oxygen that
sustains the flame. Capture all the smoke and carbon dioxide from the
burning candle. Weigh those products, along with what’s left of the candle,
and you’ll find that it all weighs slightly less than the original candle and the
oxygen. How much less? Less, in mass, by the amount E/c2, where E is the
energy released (as heat and light) by the burning candle. Because c is so big,
that’s a very small amount of mass—so small that it would be virtually
impossible to detect. Similarly, if you weigh a rubber band first when it’s
stretched and then after you release it, you’ll find it weighs less. Again, how
much less depends on how much energy you put into stretching the rubber
band. Or weigh all the uranium fuel going into a nuclear power plant and
weigh the fuel again after the plant has run for a year; you’ll find there’s less
total mass. How much less? Less by an amount E/c2, where E is the total
energy extracted from the fuel, in the form of heat and electricity. The only
difference between the nuclear power plant and the candle is that the nuclear
plant converts a greater fraction of its mass into energy. E = mc2 applies to
both processes so to claim that it’s necessary to know E = mc2 to build
nuclear weapons is to say that it was also necessary for our ancestors to
know E = mc2 before they first harnessed fire. Again: E = mc2 applies to all
processes that release or absorb energy.
Let me make this last point very clear with my nuclear power plant
example. Suppose a nuclear power plant and a coal-burning power plant
produce electricity at the same rate. Then in 1 year, each converts the same
amount of matter into energy. The only difference is that nuclear processes
are about 10 million times more effective in turning matter to energy than are
chemical reactions. For that reason the coal-burning power plant has to
process a lot more matter than the nuclear plant. A typical coal-burning plant
consumes several 100-car trainloads of coal each week, while a nuclear
plant might be refueled every year with a truckload of uranium. E = mc2
applies to both power plants; it’s just that the nuclear plant needs less fuel
because it converts a greater fraction of its fuel mass to energy.
So why is Einstein often associated with nuclear weapons? First,
because nuclear reactions are the most obvious and dramatic manifestation of
mass-energy equivalence. In fact, it’s only for nuclear reactions that reliable
measurement of mass changes is possible. Measurement of those changes, in
fact, gave physicists a new way of determining the speed of light: If you
know the energy E released in a nuclear reaction and can measure the mass
change m, then you can calculate c from E = mc2. Einstein himself speculated
as early as 1905 that newly discovered radioactive elements might contain so
much energy that they could be used to test mass-energy equivalence,
although 2 years later he ruled a nuclear test of E = mc2 “out of the question.”
A second reason for Einstein’s association with nuclear weapons is
historical. In a famous letter written in the summer of 1939, Einstein
informed President Franklin Roosevelt of the wartime potential of nuclear
energy and urged the president to ensure that the Allies did not fall behind the
Germans in nuclear weapons research. The ultimate result was the Manhattan
project that developed the world’s first nuclear weapons. But the Roosevelt
letter was not Einstein’s idea, and at the time Einstein was not particularly
familiar with developments in nuclear physics. Leo Szilard, who had first
conceived the idea of a nuclear chain reaction, visited Einstein to urge that
the great physicist lend his name to the effort to mobilize nuclear weapons
research. Einstein and Szilard jointly composed the letter, which Einstein
alone signed, and a mutual acquaintance of Szilard and Roosevelt then
delivered the message. So it was Einstein’s fame, not the substance of
relativity theory, that accounts for his only direct influence on the
development of nuclear weapons.
Back to E = mc2. I’ve emphasized that this expression of mass-energy
equivalence applies to all forms of energy. That includes the energy of
motion—what physicists call kinetic energy. The faster an object is going
relative to you, the greater its kinetic energy. Because mass and energy are
equivalent, the extra energy associated with an object’s motion manifests
itself in the same way that mass does—it makes it harder to change the
object’s motion. For example, a massive bowling ball is harder to get
moving than a much lighter tennis ball. According to Einstein, it’s even
harder to get the ball moving faster if it’s already moving relative to you.
That’s because its kinetic energy adds to the ball’s inertia, or resistance to
changes in its motion. In other words, both mass and energy have inertia.
So what? Imagine trying to get an object—for example, that bowling ball
—up to the speed of light. At first it’s not too hard to make it go faster, but as
its speed approaches a substantial fraction of c, the inertia of its kinetic
energy becomes substantial, and a given force becomes less effective at
increasing the ball’s speed. As the speed approaches c, Einstein showed, the
force and energy required to make the ball go still faster both grow rapidly
larger. As the ball’s speed approaches c, the energy you supply in attempting
to make it go even faster instead goes mostly into increasing its inertia. To
get all the way to c would require infinite force and infinite energy. That’s
simply impossible, and for that reason you can’t get a material object up to
the speed of light. You may find this argument more satisfying than my earlier
ones, because you can imagine pushing on that bowling ball and finding that
it now has so much inertia that you just can’t increase its speed very much.
But again there’s no new principle involved here; mass-energy equivalence,
like the rest of relativity, follows ultimately from the Principle of Relativity.

Tachyons, Luxons, Tardyons

All I’ve really demonstrated is that it’s not possible to get an object from rest
(relative to you) up to the speed of light. Might there be objects that move
always with speeds greater than light? Some physicists have explored that
possibility, calling hypothetical faster-than-light particles tachyons (from the
Greek “tachy” meaning “swift,” as in your car’s tachometer that measures
engine speed or the pathologically fast heartbeat termed tachycardia). If
tachyons exist, they can never be slowed down to speed c or below, and in
particular we and they can never be at rest relative to each other.
Experimental attempts to detect tachyons have failed, and many physicists
believe that they might permit causality-violating time travel into the past.
For that reason alone their possible existence is highly questionable.
It’s also possible that things other than light could travel at speed c.
Collectively, entities that move at light speed are called luxons. For decades
physicists believed that elusive particles called neutrinos belonged to this
class. Luxons must have zero mass, and for that reason their inertia doesn’t
become infinite at c. In fact, luxons can only move at c and cease to exist
when stopped. Experiments in the late 1990s showed that neutrinos almost
certainly have a very small mass, and if that’s true then they can’t move at the
speed of light. Hypothetical particles called gravitons, associated with
gravity, are predicted to go at c. For now the only certain luxons are the
photon, which in quantum physics describes the particlelike bundles of
energy that make up light and other electromagnetic waves, and the gluon, a
particle involved with the strong force that binds atomic nuclei.
Tachyons, luxons, . . . what else might there be? Well, there’s all the rest
of us material entities with greater-than-zero mass. We’re tardyons, from the
root “tardy,” for “slow” or “late.” We’re constrained to move only at speeds
less than c relative to any uniformly moving reference frame, and it’s to us
that the “cosmic speed limit” arguments of the preceding sections apply.

Quantum Weirdness

It’s not so much faster-than-light speed that relativity prohibits, rather, it’s the
transmission of information at speeds greater than c. As we saw in the
previous chapter, that prohibition is what saves the logical order of cause
and effect from the otherwise devastating fact that different observers ascribe
different time orderings to the same pairs of events. So tachyons and other
superluminal occurrences aren’t strictly banned, but if they exist they have to
do so in such a way that precludes information transmission at faster-than-
light speed.
Recently there’s been considerable interest in a strange phenomenon in
quantum physics whereby two objects are somehow “entangled” in such a
way that they seem able to communicate instantaneously even when they’re
far apart. This unusual possibility was first suggested in 1935 by Einstein
and his colleagues Boris Podolsky and Nathan Rosen. Einstein, Podolsky,
and Rosen thought up the so-called EPR paradox as a way of questioning the
probabilistic interpretation of quantum physics—an interpretation that
Einstein rejected, most famously in his claim, loosely translated, that “God
does not play dice with the Universe.” Much later, in 1964, the British
physicist John Bell showed how experiments on the EPR phenomenon could
distinguish between probabilistic and deterministic interpretations of
quantum physics. Experiments done in the 1980s and subsequently have
confirmed the probabilistic interpretation, upholding the apparently
instantaneous communication between two widely separated particles.
In a typical EPR experiment a pair of particles is created and moves
away from each other in opposite directions. The particles have certain
properties that can take on particular values; for one property those values
are described by the words “up” and “down.” It’s known for certain that if
one particle is up, then the other must be down. Now here’s the weird thing:
according to the probabilistic interpretation of quantum physics, the particles
don’t have well-defined values “up” or “down” until someone measures
them. It’s not just that those values aren’t known before measurement—it’s
that they’re not even determined. So the particles fly apart and, later,
experimenters measure one of the particles to have the value “up.” According
to quantum physics, the act of measurement is what causes the particle to
have that value; before the measurement it was undetermined. When the one
particle is measured to be up, then the other must, immediately, become
down. Indeed, experiments confirm that this is the case. Somehow the act of
measuring the up or down state of one particle in the pair has instantaneously
fixed the state of the other—even though it’s far away.
Is this superluminal communication? Are the two particles in violation of
special relativity? Physicists think not. The two particles have, in some
weird quantum-physics sense, a correlated existence. Even when separated,
they act somehow like a single entity. Perhaps this entangled pair of particles
does communicate instantaneously. Here’s the rub: there’s no way you can
use this strange entanglement to send information, because it’s entirely
random whether you measure up or down for the first particle. So you can’t
say, to paraphrase Paul Revere, “‘Up’ if by land and ‘down’ if by sea . . . ,”
because you have no way of influencing whether your measurement of the
first particle will yield up or down.
The EPR phenomenon is philosophically fascinating and reveals a
startling interconnectedness at the heart of physics. You can find out a lot
more about EPR from works on quantum physics. Here, in a book on
relativity, it’s enough to note the phenomenon and to recognize that it appears
not to violate relativity’s prohibition on transmitting information at speeds
greater than c.
Reports appear from time to time on other seemingly superluminal
phenomena, often in astrophysical situations. So far, all have proven to have
other explanations. So solidly grounded is special relativity that if you hear
of a new discovery of something moving faster than light, be skeptical! But,
like any good scientist, remain open-minded too. By the way, I need to assert
again that the cosmic speed limit is the speed of light in vacuum, c, and that it
applies to the transmission of information. It’s entirely possible for an object
to move through a material medium at a speed faster than the speed of light in
that medium. For example, the speed of light in water is about two-thirds of
c. When electrons or other particles move through water at faster than this
speed, they produce shock waves analogous to the sonic booms from
supersonic aircraft. These shock waves are essentially intense emissions of
electromagnetic radiation, produced when waves pile up on each other
because they can’t move fast enough to get away from the rapidly moving
particles. Some very high energy sources of x-rays and visible light, used in
research and technology, take advantage of this mechanism. You may also
hear of certain kinds of waves with speeds purportedly greater than c. In the
mathematical description of wave motion, there are two different speeds
associated with waves. One of them can, indeed, be greater than c—but this
speed is not associated with the transmission of any information. The second
speed, that of the information carried in the wave, is always less than or (for
electromagnetic waves in vacuum) equal to c.
The cosmic speed limit c really does appear to be inviolable!
CHAPTER 13

IS EVERYTHING RELATIVE?
• • •

It’s been fashionable to invoke Einstein in support of the idea that everything
is relative not only in physics but also in distinctly nonphysical realms like
morality, aesthetics, and politics. I’m not going to get into those
philosophical arguments here, but I do want to look closely at whether
“everything is relative” is true in physics.
Physics is the science that tells us how physical reality works. What we
might expect of physics is a truly objective description of that reality, a
description that transcends one’s own narrow point of view. Physics
shouldn’t be so much about those things that vary from one viewpoint (read
“frame of reference”) to another, but about underlying, objective truths that
everyone agrees on. That may sound lofty and grandiose, but it highlights the
reason why a physics that offered nothing more solid than shifting measures
of time and space couldn’t be altogether satisfying.
In fact, everything isn’t relative. Relativity theory itself offers a
permanence, an objectivity, that transcends the stretching of time and
squeezing of space that we find justifiably fascinating. Einstein himself was
not thrilled with emphasizing primarily what’s relative. He did not call his
own work the theory of relativity and accepted that terminology only after it
had become widely established. Further, Einstein sympathized with a group
of physicists who, in the 1920s, sought to change the name to “theory of
invariance.” But “theory of relativity” stuck, even though it describes a
theory in which, in fact, everything is not relative.
 So what isn’t relative? If such seeming absolutes as space and time turn
out to be relative to one’s frame of reference, what can be left that’s truly
absolute? The Principle of Relativity contains one answer: the laws of
physics are absolute, applying equally well in all uniformly moving frames
of reference. A corollary that I’ve stressed repeatedly gives us another
absolute: the speed of light, c, has the same measured value for all observers
in uniformly moving reference frames. Are there other absolutes, quantities
whose values don’t depend on one’s frame of reference? Yes, there are many.

Spacetime
The eminent relativity physicists Edwin Taylor and John Archibald Wheeler
begin their delightful book Spacetime Physics with the “Parable of the
Surveyors” (check out the Further Readings list for details). Here they
describe a fictitious town that is mapped by two different surveyors. One
works in the daytime and uses a magnetic compass to establish directions.
The other works at night and takes directions from the North Star. As anyone
who has worked with map and compass knows, magnetic and true north are
not quite the same; hence, the surveyors’ maps are slightly different, even
though they map the same physical town.
Suppose you want to get from point A to point B in Taylor and Wheeler’s
fictitious town. You check the daytime surveyor’s map (Figure 13.1a) and see
that you can do this by going 4 miles east and then 3 miles north. But if you
use the nighttimer’s map (Figure 13.1b) you’ll want to go nearly 5 miles east
and then just short of 1.5 miles north. Which map is right? They’re both right,
of course. Either way, you get from A to B. It’s just that one surveyor’s
definitions of north and east aren’t the same as the other surveyor’s
definitions. Each has produced an accurate map, and if you follow either
surveyor’s directions carefully, you can get to any point you want.
The daytime surveyor says, “Point B is 4 miles east of A, and 3 miles
north.” The nighttimer says, “B is nearly 5 miles east of A, and just under 1.5
miles north.” They disagree about how to describe the position of B relative
to A, but there is something both agree on: If you walk in a straight line from
A to B, rather than following compass directions, you will walk a distance of
5 miles. Figure 13.1c shows this obvious fact. On either map, the straight-
line distance is the hypotenuse of a right-angled triangle. Either surveyor can
calculate the length of that hypotenuse using the Pythagorean theorem.
Despite the fact that they’re working with different triangles whose sides
have different lengths, both get the same value—it happens to be 5 miles—
for the straight-line distance between points A and B.

Fig. 13.1 (a), (b) Two map s describing the same region but using different definitions of the comp ass directions. Instructions
for getting from A to B differ dep ending on which map one is using, but, as (c) makes clear, the straight-line distance from A to
B doesn’t dep end on one’s choice of map .

This is all very obvious, but let me make clear why it is so. It’s because
the distance between points A and B is an objective fact about the world, a
fact that does not depend on any particular choice of compass directions. A
hundred different surveyors, with a hundred different choices for the
direction of north, would come up with a hundred different maps and a
hundred different sets of instructions for getting from A to B. But all would
agree that the straight-line distance is 5 miles. In other words, the distance
from A to B is an invariant, a quantity that doesn’t depend on one’s
particular point of view as embodied, in this case, in the choice of which
direction to call north. Invariant quantities can claim to be objectively real in
a way that quantities dependent on one’s point of view cannot.
Taylor and Wheeler’s surveyor parable is an analogy for relativity. In
relativity, the points A and B are not points in space; they are events
characterized by their locations in both space and time. Spatial distances in
relativity are like east–west distances in the fictitious town. They depend on
one’s point of view—in the case of relativity, on one’s state of uniform
motion rather than one’s choice of north. Time intervals are like north–south
distances. They, too, depend on one’s point of view, that is, on one’s choice
from among the infinitely many possible uniformly moving reference frames.
The spatial distance between two events and the time between those events
are not absolute but relative to one’s frame of reference. This, of course, you
know from earlier chapters.
But everything is not relative, and there are absolutes even in relativity
theory. In Taylor and Wheeler’s parable, the different surveyors could use
their different east–west and north–south distances to calculate the same,
invariant distance between points A and B. Similarly, in relativity, observers
in different reference frames can use their different measures of spatial and
temporal separation between two events to find an invariant “distance” not in
space alone but in spacetime. All uniformly moving observers, no matter
what their states of relative motion, will agree on the value of this spacetime
interval between any two events. Because its value is independent of one’s
point of view (read “frame of reference”), that interval can claim a reality
that relativity denies to measures of space alone or time alone.
The different surveyors in Taylor and Wheeler’s town chose to break the
two-dimensional space of their town into north-south-east-west in different
ways, but the underlying reality of the town itself didn’t depend on those
arbitrary choices. Similarly, we break the underlying reality of spacetime
into space and time in different ways, depending on our states of relative
motion. It’s not space and time that have absolute, invariant reality, but rather
their combination in the single entity called spacetime. Where the surveyors’
space was two-dimensional, spacetime must have four dimensions because
there are three independent spatial dimensions and one of time.
The German mathematician Hermann Minkowski, whose classes Einstein
had attended as a student, later studied relativity and was first to conceive
the idea of four-dimensional spacetime and the invariance of the spacetime
interval. Minkowski summarized his insight in an oft-quoted statement:
“Henceforth space by itself, and time by itself, are doomed to fade away into
mere shadows, and only a union of the two will preserve an independent
reality.”*
 Given our map analogy, you might expect that Minkowski’s spacetime
interval follows from applying the Pythagorean theorem to the space and time
intervals between two events. Actually, though, the mathematics is different.
Where the Pythagorean theorem gives the square of a right triangle’s
hypotenuse as the sum of the squares of the two sides, the spacetime interval
in relativity has its square given by the difference of the squares of the time
and space intervals. That simple change from a positive to a negative sign
tells us that spacetime does not obey the rules you learned back in tenth-
grade geometry. Furthermore, it raises the possibility that the square of a
spacetime interval can be positive, negative, or zero. That mathematical
division into three classes corresponds to Chapter 11’s classification of
event pairs as being either causally related or not. For those events that are
close enough in space and far enough apart in time that an observer can be
present at both—and thus the events can be causally related— the square of
the spacetime interval is positive. Such an interval is called timelike because
an observer with just the right motion can be present at both events, so for
that observer the events are separated in time only. The numerical value of
the interval is just the time read on that observer’s clock. In other reference
frames the separation is a mix of space and time, but the time separation
always dominates; hence the name “timelike.” On the other hand, if the two
events are so far apart in space and so close in time that they can’t be
causally related, then the square of the spacetime interval is negative and no
observer can be present at both events. For an observer in just the right
reference frame, the two events will be simultaneous. For this observer their
separation is purely spatial; hence, the interval is spacelike. Take the
negative sign off its square, take the square root, and the numerical result is
the distance between the events in the frame where they’re simultaneous.
Between the general cases of timelike and spacelike intervals is the lightlike
interval, whose numerical value is zero even though the two events are
located at different places and occur at different times. Events separated by
lightlike intervals correspond to the emission and arrival of a light flash.
We’ll soon see how the unusual geometry suggested by relativity’s modified
Pythagorean theorem becomes central to Einstein’s general theory of
relativity.
The Meaning of c

There’s one further complication, and in Taylor and Wheeler’s parable it

involves an unusual religious convention. In their fictitious society, north–
south distances are considered sacred and are measured in a special unit—
the mile. East–west distances are not sacred and are measured in ordinary
meters. Different surveyors cannot consistently describe the straight-line
distance between points A and B because that distance involves two different
units for measuring spatial separation. Their religious convention keeps them
from doing the obvious—converting all distances to the same unit, be it miles
or meters. Then a brilliant and creative new surveyor arrives in town. This
newcomer compares the different surveyors’ maps, and—heresy!—converts
all distances to the same unit. Henceforth all surveyors find that they agree on
the straight-line distance between any two points. In other words, they
discover an invariant aspect of reality that doesn’t depend on their different
points of view.
Einstein is the newcomer. What his relativity shows is that time and
space, before considered separate quantities, are actually aspects of the same
underlying physical reality. That reality is obscured by the convention of
measuring time in one unit (seconds, for example) and space in another
(meters). In reality, we should use the same units for both. To do so we need
to know the conversion between seconds and meters—just as the surveyors
needed to know the conversion between miles and meters to put their
measurements in the same units. What is that conversion in the case of meters
and seconds? Simply the speed of light, c. What’s a meter of time? It’s the
time it takes light to cross a meter of space. What’s a second of space? It’s
the distance light goes in 1 second of time. We can use either unit for both
space and time, or any other distance or time unit we choose. In this book we
haven’t gone quite all the way to treating space and time with exactly the
same units. Instead, we’ve been measuring time in years and distance in
light-years. But the effect is the same; the conversion between time and space
units is the speed of light, in this case 1 light-year per year. Had we dropped
the “light-” we would have been measuring both distances and times in
precisely the same unit, namely, the year.
Viewed in this sense, the speed of light isn’t so much a speed as it is a
conversion factor between units of space and time. It’s pure coincidence that
it has the value 299 792 458 meters per second; that number is just an artifact
resulting from our arbitrary decisions to define the meter and second as we
first did—definitions that were made long before anyone knew anything
about relativity. If we embrace relativity fully, we should dispense with one
of those units and choose a single unit for both space and time. In that more
philosophically sensible system, the speed of light is just the number 1. Of
course, measuring time and space in the same units isn’t very practical for
beings who creep about their planet at speeds much less than c. A 60-mile-
per-hour highway speed limit would have to read 1/10,000,000 in those
units! But when we’re discussing hypothetical high-speed space trips, or for
physicists working with elementary particles moving through their labs at
relativistic speeds, it makes eminent sense to use units where c = 1. In those
units space and time are implicitly aspects of the same underlying reality.

What Else Isn’t Relative?

Are there other relativistic invariants besides the space-time interval? There
are, and each combines previously unrelated quantities into a single four-
dimensional entity. Here I’ll describe just a few more of these invariants,
which necessarily involve physics concepts I haven’t discussed in this book.
If you’re unfamiliar with these concepts, be content to think of the spacetime
interval as the most obvious of many quantities that are, in fact, the same for
all observers.
If you’ve studied high-school physics, you learned about momentum—the
product of an object’s mass and velocity that Newton called its “quantity of
motion.” You also know about energy, a concept I discussed in the previous
chapter. In Newtonian physics, momentum and energy are two distinct
quantities. In relativity, they’re aspects of a single four-dimensional quantity
sometimes called momenergy. Energy is the time part of momenergy and
momentum is the space part. Observers in different reference frames will
measure different values for an object’s momentum and for its energy, but
when they combine them using the same modified Pythagorean theorem I
introduced in connection with the spacetime interval, they get the same value.
In suitable units, incidentally, that value is just the object’s mass.
 Yet another invariant is related to electricity. Electric charge is a
fundamental property of matter and is an especially important concept in the
study of static electricity. After learning about static electricity, one usually
goes on to study electric current, the flow of electric charge that powers our
electrical and electronic devices. Static and current electricity may seem to
have little in common, but in fact electric charge is closely related to the time
part of a single four-dimensional charge–current entity whose space part is
related to the electric current. (It’s actually the density or concentration of
charge and current that constitutes the space and time parts of this so-called
four-dimensional current.) If you’re at rest with respect to an electric charge,
all you see are the static electric effects of that charge. If you move relative
to the charge, then it’s moving relative to you, and you see an electric current.
Different observers see different mixes of charge and current densities. If
they combine the two according to the modified Pythagorean theorem, they
get a value on which all observers agree. Once again, there’s an underlying
four-dimensional reality that’s the same for all observers. In electricity, how
that reality breaks into charge and current concentrations depends on the
observer’s frame of reference, but the underlying reality is the same for all.
While we’re on electricity, I’ll give you just one tantalizing hint of
another remarkable insight to be gained from relativity; to get the full flavor
of this one, you’ll have to dig into a more substantial physics book.† Suppose
I have some electric charges moving through a wire. For simplicity, let’s
assume positive charge is moving one way and negative charge the other,
giving a net motion of charge that constitutes an electric current (that isn’t
quite how currents in real wires work, but it will make the explanation
simpler). Assume further that the positive and negative charges are equally
spaced. Now let’s put another charge in the vicinity of the wire; we’ll call it
a test charge. If the test charge is at rest with respect to the wire, it “sees” the
wire as containing identical lines of positive and negative charge, moving in
opposite directions. According to relativity, the distance between those
charges is contracted compared to what it would be if the test charge were at
rest with respect to them instead of at rest with respect to the wire. But it’s
contracted the same amount for both positive and negative charges, since
they’re moving at the same speeds relative to the test charge (albeit in
opposite directions). So as far as the “stationary” test charge is concerned,
the wire contains equal densities of positive and negative charge, so it’s
electrically neutral. It therefore exerts no electric force on the test charge.
Now suppose that the test charge moves alongside the wire. Relative to
this “moving” test charge, the lines of positive and negative charge are
moving with different speeds. So length contraction affects each line of
charge differently and as a result the test charge no longer sees equal
densities of positive and negative charge. It “thinks” the wire is electrically
charged, and correspondingly it feels an electric force. That force is what we
usually think of as magnetism—the force resulting on a charge that moves in
the vicinity of an electric current. Viewed this way, magnetism is a necessary
consequence of electricity in a world where the relativity principle applies.
Once again, it’s relativity that unifies two seemingly separate phenomena,
in this case electricity and magnetism. What one observer sees as a purely
electric effect another will see as a combination of electric and magnetic
effects. How electromagnetic phenomena break out into separate electric and
magnetic aspects depends on one’s frame of reference. In that sense
electricity and magnetism separately, like space and time separately, have no
claim to objective reality. Underlying the relative measures of space and time
is a single, objective measure of spacetime that’s the same for all observers.
Similarly, all observers can describe the electric and magnetic phenomena
they observe as resulting from a single, underlying electromagnetic entity.
Different observers disagree about the extent to which phenomena are
electric or magnetic, but each can understand his or her observations as
particular viewpoints on an underlying, objective electromagnetic entity.

Is Everything Relative?

We now have a firm answer to the question that forms this chapter’s title.
Everything is not relative. If it were, there would be nothing absolute about
physics and that science could hardly claim to be a description of an
objective physical reality. Cherished absolutes of our common sense—in
particular, space and time—have lost their absolute status and have become
relative to our particular point of view (i.e., frame of reference). But they’re
not lost for good; they’ve simply merged into the greater, more encompassing
absolute of four-dimensional spacetime. Space and time aren’t the only
players on this four-dimensional stage; many other physical quantities that
once seemed distinct have merged into four-dimensional entities with an
absoluteness that transcends particular frames of reference. Ultimately,
relativity has enhanced, not diminished, our sense of an objective and
absolute reality.

* Hermann M inkowski, quoted in E. F. Tay lor and J. A. Wheeler, Spacetime Physics, 2d ed. (New York: W. H. Freeman, 1992),
p . 15.
† See, for examp le, R. Wolfson and J. M . Pasachoff, Physics for Scientists and Engineers, 3d ed. (Addison Wesley Longman,
1999), p p . 1035–7; the sections just before this also exp lore relativistic invariants. A more thorough analy sis is given in E. M .
Purcell, Electricity and Magnetism, 2d ed. (New York: M cGraw-Hill, 1985), chap . 5.
CHAPTER 14

A PROBLEM OF GRAVITY
• • •

What’s so special about Einstein’s special theory of relativity? Special here

doesn’t mean there’s something wonderful or noteworthy about Einstein’s
theory. Instead, special means specialized, limited, restricted—to the special
case of uniform motion. You’ve seen that the essence of special relativity lies
in a single statement: The laws of physics are the same for all observers in
uniform motion. But why this restriction? Why should it matter whether
motion is uniform? Why not the same physics for everyone, period?
These questions provide a philosophical motivation for generalizing the
special theory of relativity into a statement that drops the restriction to
uniform motion. The result would be, and is, the general theory of relativity.
Going from the special to the general theory proves more complicated—
philosophically, physically, and mathematically—than you might expect. It
took Einstein some 10 years, with many false starts and wrong turns, from his
1905 paper introducing special relativity to the completion of his general
theory in its final form. Along the way the general theory became not just a
generalization of special relativity but a whole new way of understanding
gravity.
This chapter explores the development of general relativity and shows
why it became a theory of gravity. In the last two chapters I’ll go on to
examine the implications of general relativity and its central role in modern
astrophysics and cosmology. The mathematical complexity of general
relativity means that I can’t give you the kind of ironclad logical
development that the preceding chapters provided for special relativity.
Instead, I’ll try to motivate your understanding of the key ideas in general
relativity and relate them to just a few basic principles.

A Problem with Newton

Einstein knew that Newton’s theory of gravity could not be consistent with
special relativity. In Newton’s “action at a distance” view of gravity that I
introduced in Chapter 4, Earth somehow reaches out across empty space and
pulls on the Moon, a quarter-million miles distant, exerting the force that
keeps the Moon in its circular orbit. In Newton’s view, that influence from
Earth to Moon is instantaneous. The Moon knows right now about the
presence of the distant Earth, as it is now. Suppose Earth were suddenly to
disappear. According to the action-at-a-distance view, the Moon would
immediately cease to feel Earth’s gravity and would go off on the straight-
line path that Newton says it should follow in the absence of a force. But that
implies instantaneous transmission of information from Earth to Moon, which
is inconsistent with relativity’s assertion that information cannot be
transmitted faster than the speed of light.
Another problem with Newtonian gravitation is that the gravitational
force depends on the distance between massive objects. We’ve learned that
“the distance between . . .” is not an absolute quantity; it differs from one
reference frame to another. So in whose reference frame is that distance to be
measured? And exactly when, especially if gravitating objects are in motion
relative to one another?
Incidentally, you might wonder whether electromagnetism suffers the
same problems. Is special relativity also inconsistent with electromagnetism,
with its description of the forces between electric charges and between
magnets, and of electric and magnetic fields? The answer is no. As you’ve
seen, relativity grew directly out of questions about the nature of light as an
electromagnetic wave. Special relativity developed as a theory fully
consistent with Maxwell’s equations of electromagnetism. What saves
electromagnetism is the interaction between electric and magnetic fields that
produces electromagnetic waves—waves that travel at the speed of light.
Consider an electron (negative charge) experiencing the attractive force of a
nearby proton (positive charge). In the field picture, the proton creates an
electric field in the space around it, and the electron responds to the field in
its immediate vicinity. Suppose the proton were suddenly to disappear. In
electromagnetism, news of that disappearance moves outward as an
electromagnetic wave, traveling at the speed of light. Ahead of the wave, the
electric field is just as it was before. Behind it, there is no electric field. The
region of no field expands outward, from the site of the vanished proton, at
the speed of light. Eventually the electron learns of the proton’s
disappearance as the no-field region reaches it and it ceases to feel an
electric force. Because the information gets to the electron at the speed of
light—rather than instantaneously as in the case of Newtonian gravity—there
is no contradiction with relativity. In the field picture, furthermore, a particle
responds only to the electric or magnetic field in its immediate vicinity. That
means there’s no problem with distances and times measured in different
reference frames. Electromagnetism is fully relativistic already and needs no
modification. Newtonian gravity is not relativistic, so a new, relativistic
theory of gravity is needed.

A Shaky Foundation
The foundation of special relativity is the principle that physics is the same
for all observers in uniform motion. How do you know whether you’re in
uniform motion? At first glance that seems easy. In an airplane, for example,
you can tell the difference between smooth air and turbulence. In the former
the plane’s motion is uniform; in the latter it’s anything but uniform. You
know the motion in smooth air is uniform because things around you, like the
airline peanuts you’ve spread over your seatback tray, stay put. In other
words, things in the plane obey Newton’s first law, the law of inertia: if
they’re at rest, they remain at rest. You could play tennis on this uniformly
moving plane and the tennis ball would behave as you expect—that is,
according to Newton’s laws. Turbulent air, in contrast, would send your
peanuts flying all over the place, and no way could you have a satisfactory
tennis match. So it looks like you can tell pretty easily whether you’re in
uniform motion.
We can make the criterion for deciding if you’re in uniform motion more
concise and scientific: You’re in a reference frame in uniform motion if, in
your reference frame, objects obey the law of inertia. That is, if an object at
rest in your reference frame remains at rest without the need for any forces to
hold it at rest, then your reference frame is in uniform motion. Another, and
actually better, name for such a reference frame is an inertial reference
frame, so called because of the link to the law of inertia. On the other hand,
if you need to apply forces to keep objects at rest—as you would have to do
to your airline peanuts when the plane encounters choppy air—then you know
you’re not in an inertial frame.
Or do you? How do you know for sure that there aren’t forces acting on
objects to keep them at rest? If you thought there weren’t, but there really
were, then you would mistakenly conclude that you were in an inertial
reference frame. After all, the most fundamental forces, like electricity,
magnetism, and gravity, are invisible. Having an electric force acting isn’t as
obvious as seeing your hand holding the peanuts to keep them at rest. Maybe
—just maybe—your airline peanuts are electrically charged and there’s some
clever arrangement of electric field holding them in place on the tray despite
what might, in fact, be nonuniform motion. Farfetched? Yes, but you just
might be deceived.
For electric and magnetic forces, fortunately, there’s a simple test. Not all
objects are electrically charged. So find one that is and one that isn’t. Watch
how they behave. If they behave the same way, then there’s no electric force
acting. If they behave differently, then the difference must be caused by an
electric force. The same idea works for magnetism: some objects are
magnetic and some aren’t. So simple experiments will tell you whether
electric or magnetic forces are present.
Gravity is different. Gravity affects all objects. Furthermore, as Galileo
purportedly showed by dropping different objects from the Tower of Pisa,
gravity affects all objects in the same way. That is, all objects, regardless of
how massive they are, undergo the same acceleration in response to gravity.
So gravity could be deceiving you. That is, you might think objects remain at
rest because you’re in a uniformly moving reference frame with no forces
acting, when really gravity is acting to keep them in place in a reference
frame whose motion is not uniform. Or, you might think that objects are
undergoing accelerated motion relative to you because your reference frame
is accelerating when in fact it isn’t but the objects are accelerating due to the
force of gravity. (If all this sounds a bit obscure, be patient; I’ll be giving
concrete examples in the next section.)
But surely you can tell the difference between a reference frame that is
“really” in uniform motion and one that “really” isn’t. Be careful! I cautioned
earlier about using the word “really” in cases where relativity renders it
meaningless, as in “You’re really moving and I’m really at rest.” Statements
like that are ultimately meaningless because no physical experiment can
distinguish absolutely between being at rest and moving. Only relative
motion is meaningful. That, of course, is what special relativity is all about.
Now we’re finding that there seems in principle to be no way of telling with
absolute certainty whether you’re in uniform motion. So statements like “I am
in uniform motion” are as meaningless as the statement “I am moving.” This
presents a big problem for special relativity, which is a theory about the laws
of physics being the same for observers in uniform motion. If you can’t tell
for sure that you’re in uniform motion, how can you ever know if special
relativity is valid for you?
So special relativity is on shaky ground and, again, the culprit is gravity.
The problem of instantaneous information transmission in Newtonian gravity
means that we need a new theory of gravity and a more general theory of
relativity that encompasses that new gravity. Now we have the additional
problem that gravity make us unable to tell for sure whether we’re in uniform
motion. For that reason, too, we need a relativity theory that doesn’t just
apply to uniform motion. Such a general theory must necessarily consider
gravity.

The Principle of Equivalence

My arguments in the preceding section may seem a bit abstract and
unconvincing. Now I’ll make them more concrete, following reasoning that
set Einstein on the path to general relativity.
Imagine you’re in a small, closed room at rest on Earth (picture an
elevator car). You’re holding a ball, which you proceed to drop. What
happens? The ball falls to the floor, and it does so with a constant
acceleration because of the constant force of gravity acting on it. That is, the
ball starts from rest and as it falls it goes faster and faster. On Earth, in fact,
the strength of gravity is such that the ball gains roughly 10 meters per second
of speed (about 32 feet per second) for every second that it falls. Your
dropping the ball constitutes a simple physics experiment, and it has a simple
and precise outcome: the ball accelerates downward, gaining 10 meters per
second of speed with each passing second. When I talk quantitatively about
speed and acceleration here I am, of course, talking about quantities
measured relative to the reference frame of the room. Figure 14.1a shows
this simple ball-drop experiment.
Now let’s attach a rocket motor to the room and travel to a distant region
of intergalactic space, so far from any stars, galaxies, or other objects that
gravity is truly negligible. We’ll set the rocket firing to give an acceleration
equal to 10 meters per second every second. That is, the force of the rocket
motor is making the room go faster and faster, gaining 10 meters per second
of speed in each second. You’re standing in the little room, and again you
perform the ball-drop experiment. What happens? While you’re holding the
ball, it shares your motion, and you in turn share the motion of the room.
Why? Because the floor of the room pushes on your feet, exerting a force that
gives you the same acceleration as the room. Your hand does the same to the
ball. So the room, you, and the ball are all accelerating at the same rate.

Fig. 14.1 The Princip le of Equivalence. (a) A ball released in a reference frame at rest on Earth falls to the floor with constant
acceleration. (b) Exactly the same thing hap p ens in an accelerating reference frame in intergalactic sp ace, far from any source of
gravitation.Here the acceleration is up ward and p rovided by the rocket motor. The situations of (a) and (b) are equivalent as far
as the laws of p hy sics are concerned.

Now, the instant you let go of the ball, there’s no longer any force acting
on it. So what does it do? It obeys the law of inertia, and it remains at rest
relative to a uniformly moving reference frame that had the same speed as the
room at the instant you released the ball. The room itself is not that uniformly
moving frame, since it’s accelerating. Relative to the uniformly moving
reference frame I’ve just identified, the ball remains at rest and the floor of
the room accelerates toward it at the rate of 10 meters per second in each
second. What’s this look like to you in the room? You share the room’s
motion, so it looks to you like the ball accelerates toward the floor, gaining
speed (relative to you) at the rate of 10 meters per second in each second
(Figure 14.1b). But this is just what happened when the room was at rest on
Earth! From your point of view, inside the room, the outcome of the ball-drop
experiment is the same in either situation. That experiment can’t be used to
decide whether you’re in an unaccelerated reference frame in the presence of
Earth’s gravity or, instead, in an accelerating reference frame in the absence
of gravity. Nor, in fact, can any other experiment involving the laws of
motion.
In the accelerating room, incidentally, you also feel just as you would
when standing on Earth. The floor of the room pushes up to accelerate you,
and by Newton’s third law you push down on the floor. The muscles in your
body tense and compress just as they would when standing at rest on Earth.
So your bodily sensations, too, give you no way to tell whether you’re in an
accelerating reference frame, absent gravity, or in an unaccelerated frame
with gravity present.
Now consider two other situations. First, put the little room in
intergalactic space again, but now turn off the rocket engine. What happens?
Gravity is completely negligible, so when you release the ball, it just stays
there. In this case the room is not accelerating, so it constitutes a uniformly
moving reference frame in which the law of inertia is valid. The ball was at
rest relative to this frame, and it remains at rest. It just floats there, next to
your hand (Figure 14.2a).
Finally, consider that the little room is an elevator, whose cable has
snapped. It’s plummeting toward Earth, accelerating downward at that same
10 meters per second each second. You and the ball experience exactly the
same acceleration, as shown by Galileo’s purported experiment (and by
many other more modern and very precise measurements). So when you
release the ball, relative to you in the falling elevator room, it again just
floats next to your hand (Figure 14.2b). In the two situations of Figure 14.2—
being in a truly gravity-free environment and falling freely in the presence of
gravity—the ball-drop experiment has exactly the same outcome. Within the
confines of the room, that experiment can’t help you decide which situation
you’re in. Nor could any other experiment involving the laws of motion.

Fig. 14.2 Two other equivalent situations. (a) In intergalactic sp ace, far from any source of gravity. With the rocket engine off,
the reference frame is unaccelerated. You float about the room, and a ball released from rest doesn’t go any where. (b) Exactly
the samething hap p ens in a reference frame falling freely toward Earth because y ou, the room, and the ball are all accelerating
downward at the same rate.

So the two situations of Figure 14.1 are indistinguishable, and so are the
two situations of Figure 14.2. Indistinguishable, that is, by experiments
involving the laws of motion. Why this indistinguishability? The reason,
ultimately, is that all objects experience the same acceleration in the
presence of gravity. In Figure 14.1b, the room’s acceleration mimics the
effect of gravity on any object, because all objects near Earth would fall with
the same acceleration. In Figure 14.2b, the ball floats beside your hand
because you, the ball, and the room all experience the same gravitational
acceleration.
Because no experiment with the laws of motion can distinguish them, the
two situations in Figure 14.1 are in a sense equivalent. So are the two in
Figure 14.2. What is the ultimate reason for that equivalence? It’s the fact that
gravity has exactly the same effect on all objects. If that weren’t true,
different objects in Figure 14.1a would fall with different accelerations, but
all objects in Figure 14.1b would still have the same acceleration relative to
the room—and the situations wouldn’t be equivalent. Absent other forces,
objects with dramatically different masses really do fall with exactly the
same acceleration. Apollo 15 astronaut David Scott dropped a hammer and a
feather on the Moon; with no air resistance, they hit the ground at exactly the
same time.
The simple fact of equal gravitational acceleration for all objects, known
since Galileo’s time, is nevertheless remarkably profound. Here’s why: The
mass of an object is a measure of its inertia, that is, of how hard it is to
change its motion. That’s what Newton’s second law says; it tells us that it
takes a bigger force to give a more massive object the same acceleration as a
less massive object. There’s nothing whatsoever about gravity in that
statement. On the other hand, the mass of an object is also a measure of the
force gravity exerts on that object. Drop a bowling ball and a tennis ball. It
takes a much greater force to give the bowling ball the same acceleration as
the tennis ball yet they fall with the same acceleration. Why? Because the
gravitational force on the bowling ball is greater by just the amount needed to
compensate for its greater resistance to being accelerated. In other words,
mass as a measure of resistance to changes in motion goes hand-in-hand with
mass as a measure of gravitational force. But why?
Why should an object’s inertia—the property that determines its
resistance to changes in motion—be essentially the same as the property that
determines the force of gravity on the object? Until Einstein’s time, this
essential equivalence of two different meanings of the term “mass” was
considered a coincidence, and physicists spoke of “inertial mass” and
“gravitational mass” as distinct properties of an object that happened to have
the same value. What Einstein saw, though, was a hint of a deeper relation
between gravity and accelerated motion. Einstein elevated to the status of a
fundamental principle the equivalence inherent in Figures 14.1 and 14.2, and
the underlying equivalence between the two meanings of “mass.” With the
same incisively simplifying genius he showed when he applied the principle
of special relativity to all of physics, Einstein declared the two situations of
Figure 14.1, and the two of Figure 14.2, to be fundamentally
indistinguishable by any physical experiments, not just those involving the
laws of motion. This Principle of Equivalence is at the heart of general
relativity.

Weightless!
You might object to my assertion of equivalence between the two situations
in Figure 14.2 on the grounds that the hapless elevator occupant in Figure
14.2b will smash to smithereens when the elevator hits the ground. That’s
true, but beside the point. What I’m declaring equivalent are an unaccelerated
reference frame in the true absence of gravity and a reference frame in
accelerated motion under the influence of gravity alone. When the elevator
hits the ground, strong nongravitational forces act on it and its unfortunate
occupant, and the equivalence is no more. But there’s a way around this
unhappy outcome. There is nothing sacred about falling down. The
equivalence would still hold if, in Figure 14.2b, I launched the elevator
sideways, so it fell in a curving path. Because everything in the elevator
would share its initial motion, all those objects would follow the same
curving path and would appear to float relative to the elevator. Of course the
elevator would still hit the ground, just not directly below its launch point.
But—as Newton first suggested, and as modern spaceflight confirms daily—
if I launched the elevator fast enough, and from a point above the atmosphere
to avoid the force of air resistance, it would “fall” around the Earth in the
circular path we call an orbit (recall Figure 3.4). Because all objects
experience exactly the same acceleration due to gravity, you, the ball you
attempt to drop, and everything else in this orbiting elevator would float
freely about relative to the elevator. That’s the origin of the apparent
weightlessness that occurs in spaceflight. It isn’t that there’s no gravity in
space or that objects in a spacecraft have somehow ceased to experience
gravity. It’s just that all objects experiencing only gravity have exactly the
same acceleration, so they all “fall” together and thus experience no relative
motion.
Being “weightless” has nothing, inherently, to do with being in space. It’s
just that in space one gets rid of the pesky, nongravitational force of air
resistance and the danger of hitting the ground. NASA’s “Vomit Comet”
training aircraft executes an arcing flight that essentially cancels the forces of
air resistance. Those onboard temporarily experience the same
“weightlessness” as astronauts on the Space Station. “Weightless” scenes in
the movie Apollo 13 were filmed on the Vomit Comet. Reviews praised the
film for its realism in “simulating the weightlessness of outer space.”
Nonsense! That was no simulation. Any reference frame under the influence
of gravity alone is a reference frame in which objects are “weightless.”
 The state of motion under the influence of gravity alone is often called
free fall: free because nothing other than gravity is acting; fall because
gravity is acting to change the motion of an object in the direction toward a
gravitating body like Earth or Sun. Most of us have trouble shaking the
Aristotelian view that gravity should make things move downward, as
opposed to the Newtonian view that gravity changes motion toward the
downward direction without necessarily moving an object downward. For
that reason free fall conjures up images of objects actually plummeting
toward Earth. That state is, indeed, free fall—but so is the motion of the
International Space Station, in its never-ending circular “fall” around Earth.
For that reason the term free float is a more apt description of motion under
the influence of gravity alone. The Space Station is in free float, along with
everything inside it. So is the plummeting elevator, at least for a while. Free
float simply means that the only influence acting is gravity.

Lost and Found: Uniform Motion

We’ve seen how confusion between gravity and accelerated motion makes it
impossible to know for sure whether you’re in a state of uniform motion.
That puts special relativity—valid only for uniformly moving reference
frames—on a shaky foundation.
To be specific, look again at Figure 14.2. Any experiment you do in the
free-float situation of Figure 14.2b will give exactly the same result as it
would in Figure 14.2a. We reasoned that conclusion for experiments
involving motion, and Einstein’s declaration of the Principle of Equivalence
suggests it’s true for an experiment involving any aspect of physics. So you
might think you were truly in uniform motion in the absence of gravity (Figure
14.2a) when you were really in free float, accelerating downward in the
presence of gravity (Figure 14.2b). You could be confused. That’s the bad
news.
Here’s the good news: Because the two situations of Figure 14.2 are
completely equivalent, all the laws of physics that work in the “truly”
uniform motion situation of Figure 14.2a also work in the situation of Figure
14.2b. That includes special relativity. So the free-float reference frame of
Figure 14.2b—or any other reference frame under the influence of gravity
alone—is a reference frame where special relativity is valid! We’ve lost the
ability to be absolutely sure whether we’re in uniform motion, but it no
longer matters because we’ve found a situation, namely free float, that is
logically and physically equivalent to uniform motion.
The situation of Figure 14.2a is, in fact, impossible. Gravity is never
completely gone, even in the emptiness of intergalactic space. But free float
is easy. Jump off a chair and you’re momentarily in free float. Board the
International Space Station and you’ll experience free float for as long as you
stay. In free float, physics is simple. Objects obey the law of inertia,
remaining at rest relative to the free-float reference frame unless some force
acts on them. Some force, that is, other than the Newtonian gravitational
force that acts with the same effect on everything in the free-float reference
frame. Give an object a momentary push and after that it moves in a straight
line at constant speed—relative, of course, to the free-float reference frame.
Its path relative to Earth is more complicated, but that’s beside the point.
Physics is simple in the free-float reference frame.
Furthermore, physics works exactly the same way in every free-float
frame. It doesn’t matter whether that frame is the International Space Station,
in its Earth orbit; a spacecraft on the way to Mars, as long as its engines
aren’t firing, so only gravity is acting; or a space probe orbiting the Sun or
even plunging into a black hole. The laws of physics—all the laws of
physics, according to Einstein and his equivalence principle—are just the
same in every free-float reference frame.
So in the real Universe, with gravity, free-float frames take the place of
the hypothetical uniformly moving frames that we used to build our
understanding of special relativity. Since the law of inertia holds in free-float
frames, I’ll grace every one of them with the title “inertial reference frame.” I
won’t call them “uniformly moving frames” because, at least on the face of it,
they patently aren’t. But in a real sense they’re the closest we can get to
uniformly moving frames, and they are every bit as good as “real” but
impossible states of “truly” uniform motion. Well, almost as good (more on
that shortly).
Because you can’t tell for sure whether you’re in uniform motion, the
phrase “uniformly moving reference frame” is not particularly useful. Our
newer term, intertial reference frame, is just fine. That term applies to a
hypothetical reference frame in truly uniform motion, but it also applies to
any free-float frame as well. In the real world, with gravity, the only inertial
reference frames we have are free-float frames. The special theory of
relativity applies in all inertial frames, and in the real world that means free-
float frames.

Farewell to Gravity As We Know It

Stand on Earth and you feel the tug of gravity on your body. Or do you?
Actually, what you feel are forces in your muscles as they tense or compress
to keep your body from collapsing in a heap on the ground. You weary of the
fight against gravity, so you frequently sit or lie down, letting chair cushions
and mattresses support your body.
But for real relief, get yourself into a free-float reference frame. Board
the Space Station or go skydiving (where you’re essentially in free float until
air resistance becomes significant and you approach a constant “terminal
speed”). In free float, you don’t feel gravity at all. That’s why you’re
“weightless.”
I’ve been putting “weightless” in quotes because, in the Newtonian view
of things, you aren’t weightless at all. Your “weight” means the force that
gravity exerts on you, and that force is certainly present to keep you and your
space station in orbit or to accelerate your skydiving body earthward. You
don’t feel that force because all parts of your body, and everything else in
your free-float reference frame, experience exactly the same gravitational
acceleration.
Now let’s think about how Einstein would describe the situation. The
essence of special relativity is that different inertial reference frames are
equally good places to experience the laws governing the physical Universe.
In relativity, something that has “reality” in one reference frame but not in
another can’t lay claim to truly objective reality. For example, the time
between two events—an absolute, fixed quantity in Newton’s view—
depends, in fact, on which inertial reference frame you’re in. So does the
distance between two points or the length of an object. You saw all this in the
earlier chapters on special relativity. In contrast, the spacetime interval I
introduced in Chapter 13 can claim to be an objectively real quantity, since
observers in different inertial reference frames find that it has the same
value. Unless a quantity or phenomenon can garner this universal agreement
among different reference frames, it can’t claim to be “real.”
The idea behind general relativity is to extend the Principle of Relativity
to all reference frames, inertial or not. In general relativity, a phenomenon
should be real only if observers in all reference frames agree about it. A
phenomenon that’s present in one reference frame but not another can’t be
objectively real.
That’s how it is with what you now call gravity. That heaviness, or
“force of gravity,” that you feel when standing on Earth disappears when you
hop into a different reference frame, in particular, any free-float frame.
According to Einstein, that force of gravity can’t be real because it doesn’t
appear in all reference frames. If a change of reference frame makes it
vanish, then it can’t be one of those deep, underlying, objectively real
aspects of the world. Farewell to gravity as we know it!
This is absurd, you say. Of course there’s a force of gravity! But to say
that is to contradict the now generalized Principle of Relativity. To say that
the force of gravity as you now know it is real is to say that one frame of
reference—a frame at rest with respect to Earth—is a better or more
appropriate place for making judgments about physical reality than is a free-
float frame. But the essence of relativity is to deny that any reference frame
has such special status. If something is here in one frame but not in another,
then that something can’t be real. Period.

What Is Gravity?

What you think of as gravity—the heaviness you feel standing on Earth or the
force that accelerates a plummeting stone—is not real because it disappears
in some reference frames (i.e., in any free-float frame). That doesn’t mean
there’s no such thing as gravity; rather, the real gravity must be something that
can’t be transformed away with a change of reference frame.
I’ve argued that gravity disappears completely in a free-float reference
frame, like an orbiting spacecraft or the little room of Figure 14.2b. Actually,
that conclusion isn’t quite true. Let’s make the room very much bigger, as
suggested in Figure 14.3. You, too, have become gigantic, with an arm span
of several thousand miles. You perform the ball-drop experiment again, this
time dropping not one ball but two, one from each hand. What happens? Each
ball accelerates downward, falling on a straight-line path toward Earth’s
center. Of course, you and the giant room are also accelerating downward, so
again the balls appear, at first, to float right next to your hands. Then
something strange happens. The balls begin to approach each other, the
distance between them shrinking as time goes by. You can see why this
occurs: The falling room is now so large that the direction “down” is
significantly different for the different balls. Each falls on a straight-line path
toward Earth’s center, but those paths converge, as Figure 14.3 shows. As a
result, the balls drift closer together in the free-float reference frame of the
huge, falling room.

Fig. 14.3 Tidal forces arise even in a freely falling reference frame. In the Newtonian view,these forces result from differences in
the direction or strength of gravity. Here the two balls fall toward the center of the Earth and thus their horizontal sep aration
decreases, an effect evident even to the freely falling observer. As (b) shows,this effect also squeezes the observer’s body in the
horizontal direction, while the decrease in gravity ’s strength with distance from Earth results in a stretching in the vertical
direction. I’ve assumed the falling room is sufficiently rigid that it isn’t significantly stretched or squeezed.

Here’s an effect that doesn’t go away when you jump into a free-float
reference frame. It’s an effect you can use to distinguish between the
situations of Figure 14.2. But it’s a distinctly nonlocal effect—an effect you
can’t notice unless your experience spans a large region of space. (Actually, I
should say, a large region of spacetime, because you won’t notice the effect
even in the large room of Figure 14.3 unless you let the experiment proceed
for some time.)
Newton would say that the drifting together of the two balls in Figure
14.3 is a subtle effect resulting from the variation of Newtonian gravity from
place to place, in this case a difference in the direction of the Newtonian
gravitational force at the locations of the two balls. In a related experiment,
you could imagine dropping two balls, one at your head and one near your
feet. Because Newtonian gravity is stronger closer to Earth, the lower ball
falls with a slightly greater acceleration than the upper one; as a result, the
distance between the balls would increase as they fell. A solid but
deformable object, like a blob of Jello or even your body, would experience
these effects as a squeezing sideways and a stretching vertically (Figure
14.3b). These subtle effects, known since Newton’s time, explain the
squeezing and stretching of Earth’s oceans that we call tides. For that reason,
they’re called tidal forces. In Newton’s view the tides aren’t caused by
gravity itself but by variations in gravity from place to place.
To Einstein, what Newton calls gravity doesn’t exist, because it
disappears in a free-float reference frame. Tidal effects, however, don’t
disappear, so they can claim to be real. For Einstein, those effects are the
true manifestation of gravity. Newton says, “Tidal effects result from
variations in gravity.” Einstein says, “Tidal effects result from gravity itself.”
To Einstein, gravity in the Newtonian sense doesn’t exist, so there’s nothing
to vary.

Curved Spacetime

So what is gravity? To Newton, it’s a mysterious force that somehow reaches

out across an immutable, unchanging, universal, three-dimensional space to
influence distant objects. To Einstein it’s much simpler. Gravity, Einstein
says, is synonymous with the geometry of spacetime.
Imagine yourself a tiny being confined to the surface of a large, smooth
sphere. Because you’re so tiny, your day-to-day existence occupies only a
small portion of the sphere. You study your world by doing the sorts of things
you did in tenth-grade geometry. You draw triangles, for example, and
discover that their angles add up to 180 degrees. You draw parallel lines,
follow them for some distance, and find that they don’t intersect (Figure
14.4a). You can take a journey halfway around the sphere and repeat these
geometric experiments in another small region. You’ll come to the same
conclusion—that your world obeys the laws of tenth-grade geometry. This
geometry is also called Euclidean geometry, after the mathematician Euclid,
and so you declare that your space is Euclidean.
At least, it’s locally Euclidean. The region in Figure 14.4a is so small
that you don’t notice the curvature of your world. Your measuring instruments
just aren’t sensitive enough to tell the difference between a truly flat world
and a very large sphere. That’s why Euclidean geometry works for you. (This
situation is not so far-fetched. We all know that Earth is round, but that’s
because we’re told so. Direct evidence for a round Earth is not at all obvious
from everyday life, and if we’re not airplane pilots or astronauts, most of us
could get along just fine in our local communities by treating Earth as flat.)

Fig. 14.4 (a) A small p ortion of a sp here is indistinguishable from a flat p lane and obey s the laws of Euclidean geometry.
Parallel lines never meet, and the angles of a triangle add to 180 degrees. (b) On a larger scale the sp here’s surface is decidedly
not Euclidean.Parallel lines ap p roach and would eventually intersect. The large triangle shown has three right angles, which add
to 270 degrees.

If you explore a larger portion of your sphere world, then you begin to
notice strange things. For example, you and a friend start out walking on
perfectly parallel paths. You both continue walking absolutely straight, but
after a while you notice that you’re getting closer together (Figure 14.4b).
Why? Because the straightest possible lines on the surface of a sphere, the
shortest paths between any two points, are circular arcs centered on the
sphere’s center (in geography they’re called great circles). Follow any two
of those arcs and eventually they intersect (at two points, no less, if you go
far enough). Or draw a really big triangle. I’ve shown one in Figure 14.4b
that is so big it extends from equator to pole, and I’ve made it an equilateral
triangle (all sides the same length). But its angles aren’t the 60-degree angles
of a Euclidean equilateral triangle; rather, they’re each 90 degrees, and sum
not to 180 degrees but to 270 degrees! Euclidean geometry just doesn’t hold
on this spherical world. The reason is obvious to those of us who can see the
whole sphere from afar: it’s because the sphere’s surface is curved, not flat.
Confined to the sphere’s surface, you might try to explain these strange
results by proposing a mysterious “force” that tugs on you and your friend,
pulling you away from “true” straight lines and causing your paths to
converge. Similarly, the tape measures you use to stake out that big triangle
might be deflected by the same force, accounting for the unusual angles. You
might try to learn more about this force by varying your experiments. For
example, replace your friend by an elephant and repeat the parallel-path
experiment. Obviously, the same thing happens. You and the elephant
gradually approach each other, despite the most meticulous effort at
following straight paths. That gradual approach is just the same for the
elephant as it was for your friend; therefore, you conclude that your proposed
“force” has the same effect on all objects, independent of how massive they
are. But your force idea is complicated and cumbersome, and invokes
something that just can’t be detected in a small, local region of your globe.
How much simpler is the correct explanation! The geometry of this spherical
world is not Euclidean; rather, it has curvature and that curvature accounts
fully for strange effects like the gradual approach of objects on parallel,
straight-line paths.
I’ve just presented an analogy for Einstein’s conception of gravity. We
live, says Einstein, in a four-dimensional spacetime. The geometry of
spacetime exhibits curvature, and that spacetime curvature is gravity. Not “is
a manifestation of gravity” or “is caused by gravity” or “causes gravity.” No:
spacetime curvature is gravity. Gravity is not some force that affects objects
in spacetime. Gravity is no more and no less than the curved geometry of
spacetime.
In a small, local region of spacetime—that is, in a local free-float
reference frame—you don’t notice gravity because spacetime curvature is
negligible over small regions of space and time. Similarly, a small region on
the surface of our hypothetical globe is essentially flat and obeys the laws of
Euclidean geometry. In either case, one’s world is simple when viewed
locally. On a small patch of the sphere, you can use tenth-grade Euclidean
geometry. Objects move in straight lines, parallel lines don’t intersect, and
triangles have 180 degrees. In a small, localized free-float reference frame in
spacetime, physics is simple. The law of inertia holds, with free objects
moving in straight lines at constant speed. It’s only when you go beyond your
local neighborhood that you notice deviations from simple geometry and
simple physics.
A description of physical reality in a local free-float reference frame
makes no mention of gravity. Objects at rest remain at rest, and objects in
motion remain in straight-line motion at constant speed. Make that reference
frame a space station orbiting Earth and this simple description remains true.
There’s no mention of Earth and its “gravity.” Objects in the space station
obey very simple physical laws because, locally, spacetime is flat. Here’s a
big philosophical advantage of Einstein’s view: There’s no such thing as
“action at a distance.” Rather, matter takes its “marching orders” from its
immediate vicinity, that is, from the local geometry of spacetime. When that
geometry is flat, as it always is in a small enough region of spacetime, those
marching orders say to remain at rest if you’re at rest, and to keep moving
uniformly if that’s what you’re doing. What could be simpler?
Things get more complicated only when we consider larger regions of
spacetime, large enough that spacetime curvature becomes noticeable. Then,
as in Figure 14.3 (and its analog, Figure 14.4b), we notice the effects of that
curvature on the paths of widely separated objects. We can call such effects
“gravitational,” but we could equally well call them “geometrical.” Gravity
is synonymous with the curved geometry of spacetime.
Locally, physics is always simple. The laws of physics work the same
way in every small free-float frame of reference. But because of gravity—
because of spacetime curvature—there is no one free-float frame that spans
all of spacetime. The local free-float frame in one small region of spacetime
is not the same as the local free-float frame in another region. For example, a
freely falling elevator in New York is not the same free-float reference frame
as a freely falling elevator in Bombay, on the other side of the Earth. That
physics works equally well (and simply) in both elevators is the essence of
special relativity. What general relativity does is to provide the link between
different free-float frames, giving a description of physics that is consistent
across large regions of spacetime. As we’ll soon see, general relativity also
tells how and why spacetime is curved.
Before moving on to summarize general relativity, I want to counter an
objection you might have to my sphere analogy. Aren’t there “really” straight
lines that cut through the interior of the sphere? Doesn’t the big triangle
“really” have 60-degree angles if I use those “truly straight” lines as its
sides? Yes—but I want you to consider only the surface of the sphere. That’s
a two-dimensional surface because on it you can move in only two mutually
perpendicular directions. Now, you can’t imagine that sphere without seeing
it curved in three-dimensional space. As the hypothetical tiny creature on the
sphere, though, you can do experiments that demonstrate your world’s
curvature without ever leaving its surface. Walk in a straight line and
eventually you come back to your starting point. Measure a big triangle and
find that its angles add to more than 180 degrees. You just can’t explain those
happenings in a Euclidean world. You don’t have to imagine three-
dimensional space to recognize that the geometry of your world is not
Euclid’s geometry. In fact, you can give an accurate and consistent
description of the sphere’s curved two-dimensional surface without any need
for a third dimension. The spherical surface’s curvature is an intrinsic
property of the surface itself, complete without any reference to a higher
dimension. It’s just that your three-dimensional brain finds it much easier to
acknowledge that curvature if you picture the spherical surface as being
curved in a third dimension.
Similarly, you and I have difficulty wrapping our minds around the idea
that we live in a curved four-dimensional spacetime. Four dimensions are
hard enough, and now they’re curved as well? What are they curved in, a
fifth dimension? No: just like the denizens of that sphere world, we can do
experiments in our four-dimensional spacetime that tell us its geometry is not
flat. That curved geometry is an intrinsic property of spacetime, and it
doesn’t require a fifth dimension. That curved geometry is gravity.

A Natural State of Motion

Roll a ball on that hypothetical sphere world or fly an airplane over the real
Earth on the shortest path from San Francisco to Tokyo—a path that takes you
by southern Alaska. In either case, your path isn’t a Euclidean straight line
because of your world’s curvature, but in both cases it is the straightest
possible path in that curved geometry. The ball rolling on the sphere world
just naturally follows that straightest path, and so does the plane if the pilot
doesn’t take active steps to turn it. As long as there are no forces acting on
them, both ball and plane follow the straightest possible path. Those
straightest paths in curved geometry are called geodesics.
Einstein’s law of motion for objects moving freely in the curved
geometry of four-dimensional spacetime is similar. As long as no forces act
on it, an object moves in the straightest possible path consistent with the
geometry of spacetime in its immediate vicinity. In other words, objects
follow geodesic paths in curved spacetime. That, in a nutshell, is half of
general relativity—the half that tells how objects move. This law of geodesic
motion ultimately covers everything from falling apples to planets and space
shuttles and on to the overall behavior of the Universe as a whole.
Did I say “as long as no forces act”? Was I forgetting about the force of
gravity? No: In Einstein’s view, gravity is not a force. Of course not; it’s just
the geometry of spacetime. Objects move in the simplest, straightest possible
paths when there are no forces acting. It’s just that those paths aren’t
Euclidean straight lines, because spacetime is curved. In other words,
because there is gravity.
In the broad historical context, we’ve just revisited the question I asked
way back in Chapter 3: What is the “natural” state of motion? For Aristotle,
the natural state of motion in the terrestrial realm was to be at rest as close as
possible to Earth’s center. Galileo and Newton swept away 2,000 years of
Aristotelian thinking with their assertion that the natural state is, instead,
uniform motion in a straight line. According to Galileo and Newton, uniform
motion needs no explanation; only when motion changes should we look for
a cause—that is, for a force. For Newton, one instance of changing motion is
free fall under the influence of the gravitational force. Einstein built special
relativity on the Newtonian premise, enlarging the Galilean relativity
principle inherent in the Galilean/Newtonian view to include all of physics.
Now, with general relativity, we broaden once more the idea of a natural
state of motion. In general relativity, the natural state of motion is a geodesic,
or straightest possible path, through spacetime. Such motion, which includes
Earth’s orbital motion around the Sun or the orbit of the International Space
Station, needs no explanation in terms of “force” because it’s the simplest,
most natural motion possible. Gravity isn’t a force! It’s the geometry of
spacetime! If an object isn’t moving in a Euclidean straight line, that’s to be
expected in a spacetime whose geometry is not that of Euclid.
Note once again that gravity, or spacetime curvature, is a global effect.
You’ll never observe it locally, in a small enough free-float reference frame.
Do experiments on the Space Station and the straight-line geometry of special
relativity works just fine, even as spacetime curvature guides the station on a
path whose spatial part circles the Earth. It’s only when you look at the big
picture, at how different local free-float frames are related, that you notice
spacetime curvature, or gravity.

The Twins, Revisited

General relativity says that objects in free float follow geodesics in curved
spacetime. We can use that fact to give a general relativistic description of
the twins’ star trip from Chapter 9. First I need to clear up one ambiguity.
That straightest possible path in curved spacetime need not be the shortest.
There are, in fact, two straightest possible paths between any two points on
our sphere world or on the surface of the real Earth. Both are sections of the
same circumference—a circle centered on the center of the sphere. For
instance, you can get from San Francisco to Tokyo on that straightest path that
takes you by Alaska and that’s the shortest possible route. But if you go
straight in exactly the opposite direction, you’ll also get to Tokyo. That will
be a longer way, around the whole rest of the circumference that you don’t
traverse on the Alaska route. Now, general relativity says that an object in
free-float follows the straightest possible path through spacetime but that
doesn’t mean the shortest path. In fact, a kind of “principle of cosmic
laziness” applies, in that the straightest path for a free-float observer in
spacetime is also the path on which the time between two events is the
greatest. Greatest compared with what? Compared with other observers who
also manage to move in such a way that they’re present at the two events—
observers whose motion is not free-float; that is, observers who experience
forces.
Now let’s apply this to the twins. We’ll neglect Earth’s weak gravity and
consider that the earthbound twin is essentially in free float. Then we have
two observers, the Earth twin and the space-traveling twin, who are both
present at two specific events, the departure and the arrival of the ship on its
round-trip journey. Whose clock records the longest time between these
events? General relativity gives us the answer: the clock of the free-float
twin, that is, the one who remains on Earth, in unaccelerated motion. The
other twin’s path through spacetime necessarily involves acceleration, and
therefore forces, in order to turn around and make the journey a round-trip.
So the space traveler’s clock reads less time, and therefore she returns
younger than her stay-at-home twin.
Figure 14.5 shows a Chapter-11-like spacetime diagram for the twins’
situation. On the diagram are worldlines for three different observers: the
stay-at-home twin, the traveling twin we discussed in Chapter 9, and a third
observer who takes a different round-trip journey. As with any diagram in
relativity, I should tell you the point of view (i.e., frame of reference) from
which it’s drawn. Here it’s the Earth–star frame. In that frame the stay-at-
home twin doesn’t go anywhere in space, giving a worldline that’s just a
vertical segment along the time axis. The traveling twin moves away from
Earth in space and forward in time; hence a diagonal worldline. Then she
turns around at the star, marked by the sharp bend in her worldline, and
returns to Earth. All the while she advances in time. Note that she returns to
the same place but not the same time (of course not; the journey takes time).
Remember that points on the spacetime diagram aren’t places, but events.
You might think the traveling twin’s worldline is longer, and it looks longer
on this diagram drawn on ordinary two-dimensional paper. But with the
weird negative-sign Pythagorean theorem that applies in relativity, the time
associated with her path is actually shorter than that of the stay-at-home twin.
The third observer, who goes off to the left but returns to Earth just when the
space-traveling twin does, records an intermediate time and therefore comes
back older than the traveling twin but younger than the stay-at-home twin. By
the way, you can tell from the worldline (how?) that this observer doesn’t
reverse abruptly and travels faster on return than on the outgoing trip. You
can also see that the stay-at-home twin is the only one who can be present at
both departure and return of the other observers’ spaceships without having
to undergo acceleration. For this twin, therefore, the time between the events
is the longest possible among any observers present at both events.

Fig. 14.5 Sp acetime diagram for the star trip of Chap ter 9, drawn from the Earth–star reference frame. Shown are worldlines for
three observers who are all p resent on Earth at the ship ’s dep arture and again at its return: (a) the earthbound twin, (b) the twin
who journey s to a 20-light-y ear distant star and returns, and (c) another observer who travels in the op p osite direction. Tick
marks are 10 y ears and 10 light-y ears ap art on the time and sp ace axes, resp ectively. Points mark the start and end of the
journey and the traveling twin’s turnaround at the star. Vertical gray line is the worldline of the star; Earth’s worldline is the
time axis itself.

What Curves Spacetime?

So spacetime is curved, and that curvature determines the natural, unforced

motion of free objects. How does spacetime acquire its curvature? That’s the
other half of general relativity. The first half tells how matter moves in
curved spacetime. The second half also involves matter, for it is, in fact, the
presence of matter that curves spacetime. Actually, given the relativistic
equivalence of mass and energy, it’s the presence of matter or energy that
curves spacetime.
Absent any matter, spacetime would be flat and objects would move in
the truly straight lines of Euclidean geometry. But in the presence of matter or
energy, the geometry of spacetime changes. Spacetime acquires curvature.
The so-called field equations of general relativity tell quantitatively what
curvature results from a given distribution of matter and energy. (Although I
used the words “flat” and “Euclidean” here, even in the absence of matter
spacetime still follows the weird, negative-sign Pythagorean theorem we’ve
encountered before. So it isn’t quite Euclidean, although its three spatial
dimensions do obey Euclidean geometry.)
So spacetime and matter are engaged in a sort of cosmic tango. Matter
(and its relativistic equivalent, energy) act on spacetime, giving it curvature.
Curved spacetime, in turn, acts back on matter, telling it how to behave.
Although the mathematical description of this relationship is complicated, its
essence is simple. Where matter is densely concentrated, spacetime is more
curved. When matter moves through spacetime, it does so in the straightest
possible paths. That’s it.
It’s more difficult to picture this general relativistic relation between
matter and spacetime than it is to imagine the forces and orbits of Newtonian
gravitation. But it can be done, as in our sphere example, by giving up a few
dimensions. Imagine a large sheet of rubber, stretched out horizontally. This
two-dimensional surface is an analog for four-dimensional spacetime. Right
now it’s flat. Roll a little ball along it, and the ball follows a straight line.
Now place a heavy ball on the sheet. The sheet deforms, with the ball at
the bottom of a depression caused by its own presence (Figure 14.6). Now,
you and I know the reason this happens is because old-fashioned Newtonian
gravity pulls down on the ball. But that’s not what this is about. The
important point is not that something pulls the ball down but that the presence
of the ball on the sheet, for whatever reason, distorts the sheet. The geometry
of the sheet changes from flat to curved. That’s just what happens to
spacetime in the presence of matter.

Fig. 14.6 A two-dimensional analogy for curved sp acetime. The large sp here distorts a rubber sheet. Smaller objects moving on
the sheet naturally follow curved p aths. One object is moving fast enough that it gets deflected, then continues on; the other is
trap p ed in a closed orbit around the large sp here. In both cases the motions result not from a force of attraction to the sp here
but from the curved geometry of the rubber sheet.
 Now suppose we roll a little ball along the curved sheet. Far from the
large ball, the sheet remains nearly flat, and the little ball rolls in a straight
line. As it approaches the large ball, the small one encounters substantial
curvature and is deflected from its straight-line path. In fact, you can even put
a smaller ball in “orbit” around the larger one, as shown in Figure 14.6.
Let’s make the rubber sheet perfectly transparent and look straight down
on it. All you see are the large ball and the small one. You study the motion
of the small ball and conclude that it doesn’t follow a straight line but gets
deflected toward the larger ball. That deflection is minimal far from the large
ball but becomes substantial closer in. You can even give the small ball just
the right speed that it’s deflected at a constant rate and circles endlessly
around the large ball. You might well conclude that there is an attractive
force between the two balls and that the force gets weaker with increasing
distance between them. You would, of course, be reasoning like Newton, and
discovering what you would happily call the “force of gravity.”
But those of us who can see and touch the rubber sheet know better.
There is no attractive force, no “force of gravity.” There’s only the rubber
sheet, whose geometry is curved in the presence of the large ball. The amount
of curvature depends on how close you are to the large ball. Small objects
move across this curved sheet on simple paths, namely, the straightest lines
possible in the curved geometry. This analogy, in a nutshell, describes
general relativity. There is no “force of gravity.” There’s only spacetime,
whose geometry is curved in the presence of matter. Objects move through
this curved spacetime on the simplest, straightest paths possible in the curved
geometry.
Note, by the way, that the actual curvature of the rubber sheet is not
synonymous with the curved path of the small ball. You can see that it isn’t by
looking at how the sheet is distorted. Furthermore, my rubber-sheet example
is incomplete because it shows only curvature in space, while gravity is the
curvature of spacetime. So what does an object’s path—say, Earth’s orbit
around the Sun—look like in curved spacetime? That’s impossible to draw
accurately, for the same reason you can’t make an accurate map of the
spherical Earth on a flat sheet of paper. So the picture I come up with cannot
be a realistic depiction of Earth’s motion in curved spacetime. In particular,
the orbital path I show doesn’t look at all straight or “straightest” any more
than the “straightest lines” on the globe in Figure 14.4 look straight. Those
caveats understood, take a look at Figure 14.7a. The orbital path—Earth’s
worldline, in the language of Chapter 11—is a spiral, because Earth goes
around a circle in space while advancing straight into the future in time.
Now, the diameter of Earth’s orbit is about 200 million miles—that’s 16
light-minutes, or about 30 millionths of a light-year. It takes Earth, of course,
1 year to complete 1 orbit. To show a few years’ motion, then, I’ve had to
use very different scales for the space and time measures in Figure 14.7a. If I
had used the natural relativistic units of the spacetime diagrams in Chapter 11
(i.e., light-years for space and years for time, with each unit occuping the
same physical distance on the diagram), then the time axis in Figure 14.7a
would have to be about a thousand miles long or the spatial extent shrunk to
some tens of millionths of an inch! I simply can’t make a correctly scaled
diagram without either going way off the page or shrinking the spatial scale
so small that the circular orbit becomes microscopic. (I’ve tried in Figure
14.7b, but even there the scales are nowhere near correct.) Still, you get the
point: Earth’s path through spacetime is a very, very loose spiral; our planet
advances much more in time (1 year) in each orbit than it does in space (an
orbital circumference, or 200 millionths of a light-year). Drawn to scale, that
spiral is very nearly a straight line—an indication that spacetime curvature in
Earth’s vicinity, although enough to produce our planet’s circular orbit, is just
not very great.

Fig. 14.7 Sp acetime diagrams of Earth’s orbital motion, showing two sp atial dimensions and time. Helical p aths rep resent
Earth’s worldline, its p ath in sp acetime. Time and sp ace scales are very different. In both versions the time axis sp ans 2.5
y ears, while in (a) the sp atial extent is only about 20 light-minutes. Dashed circle shows Earth’s orbit in sp ace alone. In (b) the
sp atial scale sp ans about 250 light-minutes, and Earth’s worldline shows more correctly as the nearly straight line our p lanet
follows in the modestly curved sp acetime of the Solar Sy stem.
 Curved spacetime, geodesics, matter and energy—in general relativity
we have a radically different understanding of gravity, an understanding that
arose almost single-mindedly in the brain of Albert Einstein. Philosophically,
Einstein’s theory of gravity is light-years different from Newton’s. What does
it tell us physically about our Universe? We’ll go there in the next chapter.
CHAPTER 15

INTO THE BLACK HOLE

• • •

General relativity certainly provides a philosophically different picture of

gravity, but does it tell us anything new physically? Do planets orbit, or
apples fall, differently in the general relativistic description of gravity? You
might at first think not. After all, old-fashioned Newtonian gravity does such
a good job predicting motion on Earth and throughout the Solar System that if
general relativity’s predictions were significantly different they would have
to be wrong!
In fact, the predictions of general relativity do differ from those of
Newton. On Earth, and indeed everywhere in our Solar System, those
differences are so slight as to be almost immeasurable. Almost, but not quite.
Very sensitive tests can distinguish Einstein’s predictions from Newton’s.
Newtonian gravitation works fine for everyday tasks like launching a space
shuttle, sending a spacecraft to Jupiter, or predicting eclipses. But when the
utmost precision is needed, and when instruments are capable of measuring
with great precision, then we can and sometimes must use general relativity
to predict the effects of gravity.
Einstein and Newton differ only slightly in regions where gravity is
weak. I’ll give a precise definition of what I mean by “weak gravity” and
“strong gravity” later in this chapter. Suffice it to say for now that nowhere in
our Solar System—even in the vicinity of the Sun itself—is gravity at all
strong. The same is true around most ordinary stars and planets that orbit
them, but there are other places in the Universe where gravity is truly strong.
There, Newtonian gravitation fails dramatically and the predictions of
general relativity once again play havoc with our commonsense notions of
time and space.

Errant Orbits
By the nineteenth century, Newtonian gravitation reigned supreme throughout
the Solar System. Meticulous calculations of the gravitational effect not only
of the Sun but also of the other planets had accounted for nearly every detail
of the observed motions of the planets. But there was one tiny discrepancy in
the motion of the planet Mercury, a discrepancy that remained even when the
effects of the other known planets were included.
In an idealized Newtonian universe containing just the Sun and one
planet, the planet’s orbit would be a perfect ellipse that closes back on itself
and repeats exactly forever (Figure 3.2 showed such an orbit). The real Solar
System is more complicated, largely because of the gravitational effects of
the planets on each other. We can subtract out those effects and ask if, in their
absence, a planet’s orbit would be the ideal, closed ellipse. For Mercury, the
answer is “not quite.” Mercury’s ellipse doesn’t quite close, meaning that the
long axis of the ellipse rotates slowly with time, as shown in Figure 15.1.
“Slowly” is the operative word here, for in 100 years the orbit swings
through a mere 43 seconds of angle. You probably don’t have a good feel for
a “second of angle,” but you know a “degree of angle” as a very small angle
indeed; there are 360 degrees in a full circle, 90 in a right angle. One second
of angle is a minuscule 1/3,600 of a degree. So the orientation of Mercury’s
orbit changes by only about 43/3,600 of a degree in a century, or just over a
ten-thousandth of a degree per year! That’s a pretty small angle, and its
measurement is all the more impressive because Mercury’s orbit is not the
obvious ellipse of Figure 15.1, but is nearly circular. Nevertheless,
astronomical observations are accurate enough that nineteenth-century
astronomers were confident in their measure of Mercury’s anomalous orbital
motion. They called the phenomenon “the precession of the perihelion of
Mercury,” where precession is a slow rotation like the change in the
orientation of a spinning top and perihelion is the point on the orbit where
the planet is closest to the Sun. You can see from Figure 15.1 that precession
of the orbital axis carries with it that point of closest approach.
Fig. 15.1 Orbital p recession. In Newtonian gravitation, a p lanet’s ellip tical orbit would rep eat forever (thick ellip se),but general
relativity p redicts that the orientation of the orbit should change with time—an effect observed in M ercury ’s orbit.The
p recession shown here is highly exaggerated, and M ercury ’s orbit is also much closer to circular.

Mercury’s perihelion precession posed a quandary for nineteenth-century

physicists and astronomers. Because it was such a small effect and because
Newtonian gravitation was so spectacularly successful, most believed it
could be explained in the Newtonian context as the effect of an as yet
undiscovered planet, an unseen moon of Mercury itself, or a swarm of
asteroids. A few tried to modify Newton’s law of gravity, but without much
success or theoretical justification.
Then, in 1915, as he was putting the finishing touches on his general
theory of relativity, Einstein tackled the question of Mercury’s orbit. To his
profound delight, he found that his theory predicted a perihelion precession
of 43 seconds of angle per century! This result was a brilliant confirmation
of what was otherwise an abstract, mathematical exercise in formulating
Einstein’s new description of gravity as the curvature of spacetime.
So one difference between Einstein’s and Newton’s theories is that, in
Einstein’s theory, planetary orbits don’t quite close. This effect is hardly
noticeable in the weak gravity of our Solar System, and it’s no wonder it was
discovered in Mercury—closest to the Sun and thus the planet subject to the
strongest (but still very weak) gravity. Philosophically, the explanation of
Mercury’s perihelion precession in terms of general relativity represents a
profound leap in our understanding of gravity. Physically, though, it’s hardly
a big deal, again, because gravity in our Solar System is so weak.

An Astrophysical Interlude
Observed differences between Newtonian gravitation and general relativity
remained mostly small, marginal effects through the first two-thirds of the
twentieth century. We’ll soon look at some of the other classical tests that
affirmed Einstein’s theory using very sensitive measurements on Earth or
elsewhere in our Solar System. But in the 1960s things began to change.
Astrophysicists discovered amazing objects whose gravity was really strong
in the sense I’ll define shortly. The first such objects discovered were stars
that, having expended the nuclear fuel that kept them shining, had collapsed
under the influence of their own gravity to form one of two remarkable stellar
endpoints. Neutron stars, the somewhat less bizarre of these two “dead star”
possibilities, pack roughly the mass of the Sun into a region only a few miles
across. Under the resulting intense gravity, electrons and protons collapse
together to make neutrons, hence “neutron star.” The neutron star is somewhat
like a giant atomic nucleus, so dense that a teaspoon of neutron-star matter
would weigh more than 100 million tons. An unusual kind of pressure arising
from all those neutrons keeps the star from collapsing further, but there’s an
upper limit to the mass neutrons can support. Above this limit, the burned-out
star must collapse completely, forming a black hole. Much more on black
holes later—they’re in this chapter’s title—but for now suffice it to say that
massive stars collapse at the ends of their lives to form either neutron stars
or black holes. In both cases, gravity in the immediate vicinity of the
collapsed star is most definitely strong. (By the way, our Sun is not
sufficiently massive for such a dramatic fate; it will eventually become a
white dwarf star, with its mass packed into roughly the size of the Earth.)
It was in the 1930s that physicists first used general relativity to predict
the theoretical possibility of neutron stars and black holes, but it wasn’t until
the 1960s that these remarkable objects were actually discovered. I won’t go
into the details of how we detect neutron stars and black holes, except to note
that they often occur in binary-star systems and are observed through their
effect on a companion star. (You may recall from Chapter 5, where
observations of binary stars showed that the speed of light does not depend
on the speed of its source, that about half the stars in our galaxy are in binary
systems.) In 1974 astrophysicists Joseph Taylor and Russell Hulse made a
remarkable discovery: a binary system containing not one but two neutron
stars in very close orbits. Each has about 50 percent more mass than our Sun
and their orbits are decidedly elliptical. At closest approach, they’re only
about half the Sun’s diameter apart. With such tight orbits, each completes a
full orbit of the other in a mere 8 days. The gravity each neutron star
experiences is much stronger than anything in our Solar System, and the
effects of general relativity are much more pronounced. Thus this unusual
binary system provides a laboratory for studying general relativity.
How do we know so much about the Taylor–Hulse binary system?
Because neutron stars, in addition to being dense, often rotate at high speeds.
This rapid rotation occurs for the same reason that a figure skater or ballet
dancer goes into a rapid spin by pulling in his or her arms. As a large star
collapses to a neutron star, its initially slow spin gets amplified as the matter
of the star “pulls in” to occupy a much smaller space. (The physics term for
this is “conservation of angular momentum.”) Furthermore, neutron stars have
strong magnetic fields, and these fields channel outgoing electromagnetic
radiation—light, radio waves, x-rays—into narrow beams that sweep
through the cosmos as the star rotates. When a beam sweeps by Earth, our
telescopes or x-ray detectors record a pulse of radiation. As the neutron star
spins, we get one such pulse for each rotation; for that reason we call these
spinning neutron stars pulsars. By timing the pulses, we can measure the spin
rate with amazing accuracy; in the Taylor–Hulse system, for example, the
spin rate of one neutron star was measured to be 16.940539184253 rotations
per second. (This measurement was made in 1986, and the rate has changed
since then—more on this later!)
Now, as I noted in Chapter 5, the motion of stars in a binary system
toward or away from us changes the frequency or color (but not the speed!)
of the light we receive. Similarly, the pulse rate we measure depends on
whether a pulsar is moving toward or away from us. (This phenomenon,
called the Doppler effect, also occurs for sound and is probably familiar to
you. For example, when you stand near a highway, you hear a high-pitched
sound as a truck approaches, then a low-pitched sound as the truck passes
and heads away from you: “aaaaaaaaaeeeiiiooooooooo.” The high pitch
occurs because wave crests pass you more often as the truck approaches; the
low pitch because they pass less often as the truck recedes. The same idea
holds for light-wave crests or pulsar pulses.) So by timing a pulsar’s pulses,
we can keep track of its orbital motion. Even though the system is way too
distant for us to see the individual neutron stars, pulse timing allows us to
construct a picture of the orbital motion. Because the pulse rate is known
with such precision, that picture is very accurate.
What’s all this got to do with general relativity? With objects as massive
and as close as those in the binary pulsar, general relativity predicts a
precession of the orbital axis by more than 4 degrees per year, some 35,000
times Mercury’s paltry 43 seconds of angle per century. Given our accurate
picture of the pulsar’s orbit, derived from timing its pulses, the orbital
precession is obvious and easy to measure. The result, from more than 20
years of observations, is a precession that’s right in line with the general
relativistic prediction.
So here’s one way in which general relativity and Newtonian gravitation
differ: elliptical orbits don’t quite close, but precess at a rate that depends on
how strong gravity is at the location of the orbit. In our Solar System, gravity
is so weak that this precession is barely noticeable. Elsewhere in the
cosmos, though, nature provides remarkable systems—here the binary pulsar
—where general relativity’s deviation from Newtonian physics is much more
obvious. That will be the story of the rest of this chapter: we identify a
general relativistic effect and confirm it early on with a very sensitive
terrestrial or Solar System–based experiment. Later, toward the end of the
twentieth century, astrophysics provides much more dramatic confirmation of
the same effect. Today, in the twenty-first century, general relativity is a
solidly verified theory, a working tool of many astrophysicists.

Light Bends!

In the previous chapter I introduced the Principle of Equivalence as the heart

of general relativity. This principle says that no experiment can distinguish
the state of free float—motion under the influence of gravity alone—from
truly uniform motion in the complete absence of gravity. Figure 14.2 made
that point very clear for a simple mechanical experiment like dropping a ball.
Einstein extended the equivalence principle to all of physics, including
electromagnetism and the behavior of light. He then turned his reasoning
around and used the principle to deduce a new physical phenomenon, namely,
the bending of light in the presence of gravity.
Consider once again the little room of Figure 14.2a, way off in
intergalactic space where there’s no gravity whatsoever. Suppose we mount
a flashlight on one wall and shine the light directly across the room. The light
goes straight across the room to make a bright spot on the opposite wall,
directly opposite the flashlight (Figure 15.2a). Now switch to the situation of
Figure 14.2b, where the room was shown falling freely toward Earth.
Einstein, with his Principle of Equivalence, says that this situation is
indistinguishable from that of Figures 14.2a and 15.2a. That means any
experiment we do in the falling room must have exactly the same outcome
that it did way off in intergalactic space. If it didn’t, we would have a way to
distinguish the two situations. So if we turn on the flashlight here, the light
must still hit the wall directly opposite. In either situation, an observer in the
room sees the light go through the room in a straight path to the opposite
wall. Now think about what this looks like to someone on Earth watching the
falling room. To the earthbound observer, light leaves the flashlight at the left
side of the room. By the time it gets to the opposite wall, the room has fallen
some distance, yet the light still hits the wall right opposite its source. How
can that be? Only if the light, too, has “fallen.” That is, the light itself must
describe a curved path, as suggested in Figure 15.2b.
Fig. 15.2 (a) Light shined across the little room in gravity -free sp ace hits the op p osite wall. (b) The freely falling room is
equivalent, so here the light should hit the op p osite wall the same distance below the ceiling as it did in (a). But the room is
accelerating downward,so the light—s p ath must bend. The room is shown twice, first when the light dep arts the left wall and
again (lighter image) when it hits the right wall.

Einstein’s conclusion that gravity must affect light shows the power of the
equivalence principle in deducing new physical phenomena. Later, when he
had formulated the full general theory, Einstein had another explanation:
light, like matter, moves in the straightest possible paths through spacetime.
But because spacetime is curved near massive objects, those paths are not
the true straight lines of tenth-grade geometry. Incidentally, Einstein’s final
form of general relativity predicts a bending of light twice as great as his
earlier equivalence-principle argument would imply.
So the path of a light ray should bend as the light passes a massive body.
To test this idea, astronomers sought to measure the bending of light passing
close to the most massive body in our Solar System—the Sun. The idea was
to view one or more stars at a time when their light had to pass close to the
Sun on its way to Earth (Figure 15.3). Astronomers could then measure
changes in the stars’ apparent positions as compared with observations made
when the Sun was not in the same general direction as the stars. From those
changes they could calculate the angle by which the Sun’s gravity had
deflected the light.
There’s a problem with this approach: To view stars that appear in the
sky near the Sun, one has to observe the stars in the daytime, but stars aren’t
visible in the bright daytime sky. So the only way to make this observation is
during a total eclipse of the Sun. Ideally, one would like an eclipse that
occurred at a time when many bright stars would appear near the Sun. Now,
Einstein made his final calculation of the bending of light in 1916.
Fortunately, an eclipse was to occur on May 29, 1919, and on that date it
happens that many bright stars appear near the Sun. So the 1919 eclipse
would provide an excellent test of Einstein’s prediction.

Fig. 15.3 (a) Light from a distant star comes directly to Earth. (b) When the Sun and star are in the same region of the sky, the
Sun’s gravity deflects the starlight. Observers on Earth then see the star at a different ap p arent p osition in the sky. Dashedlines
mark the directions observers would look to see the star. The deflection shown is greatly exaggerated.

Sir Arthur Stanley Eddington, a prominent British astronomer and early

student of Einstein’s work, together with colleagues began planning an
expedition to the 1919 eclipse. Unfortunately World War I intervened.
Eddington, a devout Quaker, sought exemption as a conscientious objector to
military service. After several appeals, his exemption was finally granted, in
part on the grounds that he was essential to the upcoming eclipse expedition.
After the war’s end, the British mounted two expeditions to view the 1919
eclipse, one to Brazil and the other, with Eddington its leader, to Principe
Island off the African coast. After careful analysis of the observations,
Eddington reported the mathematical details to a joint meeting of the Royal
Society and the Royal Astronomical Society: both eclipse sites confirmed
Einstein’s prediction. It was this result, not the special theory of relativity or
the Mercury perihelion precession, that catapulted Einstein to fame. The
London Times trumpeted the news with the headline “Revolution in Science .
. . Newtonian Ideas Overthrown.” Einstein spent the rest of his life in the
public spotlight.

Gravitational Lenses

Like the precession of Mercury’s perihelion, the bending of light remained

through most of the twentieth century an obscure and very subtle effect
significant only in helping confirm the general theory of relativity. Einstein
himself saw greater possibilities, suggesting that massive bodies deep in
space might provide much more dramatic bending of light—enough to
produce distorted and even multiple images of more distant objects. Figure
15.4 shows how this works. Here, a massive galaxy lies between Earth and a
quasar. Quasars are distant objects so bright they can outshine an entire
galaxy, yet small enough that they appear as pointlike sources of light (more
on quasars later). Light from the quasar bends so much in passing the galaxy
that an observer on Earth can see the quasar by looking in either of two
directions—above or below the galaxy in Figure 15.4. The result is two
images of the same object! In a telescopic photograph, these appear as two
distinct objects at different positions on the photo. Actually, things get more
complicated; in three dimensions the situation of Figure 15.4 would result in
the quasar’s image being smeared into a circular ring, provided the quasar
were directly behind the galaxy and the galaxy perfectly circular. Absent this
perfectly symmetric situation, the usual result would be several images of the
same quasar.

Fig. 15.4 Gravitational lensing. A massive galaxy lies between Earth and a distant quasar. Light from the quasar takes several
p aths around the galaxy, resulting in multip le images of the quasar as viewed from Earth. Dashed lines show two different
directions in which the quasar is visible, resulting in two distinct quasar images. Additional images,or a continuous ring, would
ap p ear in three dimensions.
 It was not until the 1970s that astronomers firmly identified such
gravitational lenses. By now, numerous examples of galaxies lensing
quasars have been discovered. One of the best known is the “Einstein cross,”
a constellation of four images of the same quasar, shown in Figure 15.5a. In
other examples, distant galaxies appear smeared out and distorted as their
light passes a nearer galaxy (Figure 15.5b). Study of these images gives
information about the lensing object as well as the distant galaxies.

Fig. 15.5 Hubble Sp ace Telescop e images showing gravitational lensing. (a) The Einstein Cross includes four images of the same
quasar, gravitationally lensed by a massive galaxy. The central bright sp ot is the nucleus of the galaxy, with the four quasar
images around it. Image quality is limited by the resolution of the Hubble Telescop e. (b) Here an entire cluster of galaxies acts
as a gravitational lens for more distant galaxies, p roducing distorted, arclike images. Some of these are multip le images of the
same distant galaxy.The lensing galaxies are visible as the larger, brighter objects. (Credits: (a) NASA, J.Westp hal, W. Keel; (b)
NASA, A. Fruchter, and the ERO team [STScI, ST-ECF].)

Today, gravitational lensing is more than a novel phenomenon that

confirms Einstein’s remarkable insight. It’s also become a tool for exploring
the cosmos. In Chapter 1, I described how a massive galaxy cluster some 2
billion light-years from Earth acts as a giant telescope, concentrating the light
from more distant objects that would otherwise be too dim for us to detect
with human-made telescopes. Lensing also gives astrophysicists insight into
the overall structure and evolution of the Universe by providing distances to
remote, multi-imaged quasars and estimates of the cosmic acceleration (more
on this in the final chapter). Closer to home, gravitational lensing occurs in
our own galaxy when a massive object passes in front of a star. As seen from
Earth, the star appears to brighten briefly due to the focusing of its light by
the object’s gravity. Such microlensing has become an important tool for
astronomers searching for the “missing mass” known to comprise much of the
matter in the Universe. The technique might also help astronomers discover
extrasolar planets by the microlensing effect as an unseen planet passes in
front of its star.
In describing gravitational lenses, I’ve been using language like “light
bends” and “focusing of light by gravity.” That language helps explain the
phenomenon, but it isn’t quite relativistically correct. What’s actually
happening, of course, is that light is going in the straightest possible lines in
curved spacetime. It’s just that those aren’t the ideal straight lines of tenth-
grade geometry. Sometimes—as when multiple images form—the geometry
of spacetime is such that there’s more than one “straightest line” between a
distant object and Earth.

Warping Time

In general relativity, gravity is the geometry of spacetime. So it should come

as no surprise that time itself is affected by the presence of massive objects
that warp spacetime. Once again it was the equivalence principle that led
Einstein to this new phenomenon. He imagined the situation of Figure 15.2,
but this time with the flashlight shining from floor to ceiling. In the frame of
Figure 15.2a, with gravity absent, the light reaches the ceiling unchanged; in
particular, it’s the same color as it was at its source. According to the
equivalence principle, that must also be the case in the equivalent freely
falling frame of Figure 15.2b. To an observer on Earth, though, the room is
accelerating downward as the light makes its journey from floor to ceiling.
So when the light reaches the ceiling, the room is moving faster than it was
when the light was emitted. That means an observer at the ceiling should see
a Doppler shift, toward higher frequency and bluer color. But the
equivalence principle says that doesn’t happen, so there must be a redshift,
associated with the presence of gravity, that cancels the blueshift. Observers
who aren’t accelerating downward would still see this redshift even though
there would be no blueshift. So if I’m above Earth or some other gravitating
body, and I observe light emitted at that body, then when I receive the light
it’s redshifted—that is, it has a lower frequency—compared to what an
observer right at the light source would see. Gravitational redshift is one
name for this phenomenon, but a deeper name is gravitational time dilation.
 Why “time dilation”? Because the vibrations of a light wave, like any
other periodically repeating phenomenon, provide a measure of time. In fact,
our most accurate clocks—the atomic clocks that establish world time
standards—use as their ticks the frequency of light waves emitted by
particular atoms. So in the frequency and wavelength of light waves we have
a measure of time itself. If I’m looking down at a source of light, the light
reaching me has a lower frequency than it would if I were right beside the
source. That is, the time interval between crests of the light wave is for me
stretched out. But light frequency is simply a measure of the underlying
passage of time; therefore, time intervals between any events occurring down
at the light source also appear to me stretched out. That includes the interval
between ticks of an ordinary clock; in other words, if I look down on a clock,
I see it running slow compared to a clock right next to me. That’s
gravitational time dilation. As with the time-dilation phenomenon we
encountered in special relativity, gravitational time-dilation isn’t about light
or clocks—it’s about time itself. Time runs at different rates depending on
where you are in relation to a massive gravitating object.
Is the lower clock really running slow? If I go down there and stand next
to it, its timekeeping will seem perfectly normal. Of course; the clock is in a
perfectly legitimate frame of reference and so it works normally. But
compared with the clocks of an observer located higher up, the lower clock
really does run slow. This time-dilation effect, unlike the time dilation of
special relativity, is not reciprocal. That is, a clock lower down really does
keep time more slowly compared with one higher up, and the clock higher up
really is running fast compared with the lower clock.
Does gravitational time dilation actually happen? It does and it’s been
measured. Here on Earth the effect, like all manifestations of general
relativity, is subtle and difficult to detect. However, in a famous 1960
experiment, physicists managed to measure the difference in timekeeping
rates of clocks at the top and bottom of a 74-foot tower at Harvard
University. Their clocks were atomic nuclei emitting radiation whose
changes in wavelength could be detected with exquisite precision. The result
—a shift of only about a thousandth of a trillionth of the original wavelength
—verified general relativistic time dilation for the weak gravity at Earth’s
surface. In the 1971 experiment I introduced in Chapter 1, scientists flew
atomic clocks around the world and compared their timekeeping with an
atomic clock left behind. The combination of special relativistic time
dilation associated with relative motion and reduced gravitational time
dilation from higher altitude were fully consistent with relativity. (See
“Around the World Atomic Clocks,” in the Further Readings, for details of
this experiment.) Today, the Global Positioning System (GPS) times signals
from a constellation of orbiting satellites to provide precise locations
anywhere on Earth. So accurate is GPS that if the satellites’ atomic clock
times weren’t corrected for gravitational time dilation, the system would
soon be off by a matter of miles!
As always with general relativistic effects, it’s in the astrophysical realm
that gravitational time dilation is most obvious. Even the Sun’s weak gravity
produces a measurable effect. Decades before the earthbound Harvard
experiment, astronomers had measured the gravitational redshift of light from
a white dwarf star, which boasts the Sun’s mass crammed into the size of the
Earth. At the surface of such a dense object, gravitational time dilation is
some 30 times greater than that of the Sun. Finally, today’s astrophysicists
routinely observe substantial time dilation in the strong gravity around
neutron stars and their even more bizarre cousins, the black holes.

Black Holes

What goes up must come down, right? No! Throw a ball straight up as hard
as you can and it eventually slows, stops, and returns. If you could throw it
fast enough—for an object thrown from Earth’s surface, “fast enough” is
about 7 miles per second—the ball would have enough energy to escape
Earth’s gravity altogether and would travel outward forever without
stopping. That’s because gravity weakens so rapidly with increasing distance
that escape to an infinitely great distance does not require infinite energy. The
7 miles per second you’d need is called the escape speed for Earth’s surface.
Although the human arm can’t propel anything at escape speed, rockets can.
Spacecraft traveling to the outer planets, for example, leave Earth’s vicinity
at greater than escape speed. Pioneer and Voyager spacecraft even exceed
escape speed for the Sun, meaning they’ll eventually leave our Solar System
and spend the foreseeable future drifting through the galaxy.
 What determines escape speed from a given location? Ultimately, it’s the
strength of gravity at that point. For the surface of a planet or star, that’s set
by the mass and size of the object. Cram more mass into a given-size object
and escape speed goes up. Shrink an object of fixed mass and again escape
speed goes up. So imagine compressing Earth to ever-smaller sizes. Escape
speed from the surface of the shrinking planet rises from its current 7 miles a
second to ever-higher values. It gets harder and harder to “throw” a
spacecraft or other object forever outward, but with sufficient energy and
advanced technology, it remains possible. Possible, that is, until escape
speed reaches the ultimate value, namely, the speed of light. For Earth, that
would happen when the entire planet was a little under an inch in diameter.
That’s right—you, me, Mount Everest, New York City, all the water in the
oceans, all the continents, the liquid and solid cores of the planet—all
crammed into a space smaller than a Ping-Pong ball. If this seemingly
impossible compression occurred, then we would have an object so dense
that not even light could escape. That’s a black hole.
Our incredible shrinking Earth, compacting until its escape speed
approaches the speed of light, finally provides a solid definition of the terms
“strong” and “weak” gravity that I’ve been using throughout this chapter.
Strong gravity exists where escape speed is an appreciable fraction of the
speed of light, c. Weak gravity means escape speed is far less than c. Earth’s
and Sun’s escape speeds, at 7 and 380 miles per second, respectively, are far
less than the 186,000-mile-per-second speed of light. Gravity everywhere in
our Solar System is weak. At the surface of a typical neutron star, however,
escape speed is about two-thirds that of light. This is strong gravity! Shrink
that neutron star even a little bit and it will collapse to a black hole with
escape speed c—the ultimate in strong gravity.
A black hole is a remarkable object. Our current understanding of
physics suggests that once an object has been squeezed to black-hole size,
there’s no force in the Universe that can prevent its further collapse to a
single point of infinite density. This infinite conclusion may change somewhat
when we finally learn how to merge general relativity with quantum physics,
the theory that describes matter at the atomic and subatomic scales. Even so,
black holes will remain objects in which matter is compressed to a near
point of incredible density. Surrounding this point is a spherical surface
called the event horizon, which bounds the region within which the escape
speed exceeds the speed of light. No light can escape from within this region,
making it a true horizon. Those of us on the outside can never see in, past the
horizon. There’s simply no way for us to get information about events
occurring within the horizon, hence the name event horizon.
Because light can’t escape a black hole, and since no material object can
go faster than light, that means nothing whatsoever can escape the hole. That
fact makes black holes remarkably simple objects. From the outside, black
holes exhibit very few distinguishing properties. Most significant is their
mass—the total amount of matter and energy that has fallen across the
horizon. It doesn’t matter whether that mass-energy was in the form of stars,
planets, people, interstellar dust, mice, water, light, or whatever. Once it’s
across the horizon, we can’t know anything about it and so all that matters is
the total mass. That mass determines the size of the event horizon and the
gravitational influence the hole has on the surrounding Universe. If the
infalling matter has electric charge, the black hole, too, will be charged, and
the charge will be felt outside the horizon. If the infalling matter has
rotational motion, the black hole will itself be spinning in a way that
influences spacetime outside the horizon. But that’s it: mass, spin, and
electric charge are the only properties that distinguish black holes.
People often picture a black hole as sucking up all the matter in its
vicinity. That’s a misconception, because a black hole’s gravitational
influence is the same as that of any other object with the same mass. Far from
the hole, matter will orbit in essentially Newtonian elliptical orbits
determined by the hole’s mass alone. If Earth suddenly collapsed to a black
hole, for example, the Moon would be completely unaffected and would
continue in its orbit about the Earth-mass black hole. It would not suddenly
be sucked in any more than the Moon or a satellite is sucked to Earth by the
planet’s gravity. The only objects that strike Earth are those that are on a
collision course with our planet or are close enough that Earth’s gravity
deflects them toward a collision. The same is true with a black hole; only
matter that comes very close to the event horizon actually falls through it and
because typical event horizons are very small, such a course is quite
improbable. That means an isolated black hole will swallow matter at a
rather low rate. On the other hand, a hole surrounded by a dense aggregation
of matter—as in a binary star system or near a galactic center—will generate
a substantial inflow of matter. More on this when we consider real black
holes out there in the cosmos.
The notion of a black hole behaving like a cosmic vacuum cleaner does,
however, have some merit. That’s because the event horizon is a one-way
street; matter that crosses the horizon can never re-emerge. So a black hole
only grows in mass. Again, it doesn’t do so by inexorably pulling in
everything around it; rather, whatever happens to fall into the hole simply
doesn’t get out.
Actually, even that conclusion has to be tempered. In a remarkable
quantum-physics process first envisioned by Stephen Hawking, black holes
can actually lose mass by evaporation involving particles created in the
vacuum just outside the event horizon. For astronomical-size black holes, this
process is so feebly slow as to be completely negligible, but it might play a
role in the very long-term evolution of the Universe.

Journey into a Black Hole

Black holes provide the ultimate time warp. That’s because gravitational
time dilation becomes infinite at the event horizon of a black hole. To see
what this means, imagine that you and I are positioned a fixed distance far
from a black hole. Each of us has a clock, and we’ve carefully synchronized
them to read the same time. You then proceed to fall toward the black hole
while I remain behind. As you fall, gravitational time dilation slows your
clock relative to mine. So I see your clock, and indeed all manifestations of
time, running slower and slower as you approach the event horizon. The
movements of your body appear to slow. If I monitor your heartbeat, it slows.
You appear to age more slowly. Time itself is running slower for you, in the
warped spacetime around the hole, than it is for me. As you approach the
event horizon, I see your time running ever slower until, at the horizon itself,
it stops completely! I never see you quite reach the horizon because by the
time that happens, it’s the infinite future for me! So if I’m watching matter
falling into a black hole, it appears to me to “freeze” just outside the event
horizon. In my time frame, infalling matter never crosses the horizon. For that
reason Russian astrophysicists coined the term “frozen star” to describe
black holes forming from collapsed stars.
 By the way, if you decide to abort your black-hole journey and blast
away using a powerful rocket, you can of course do so as long as you haven’t
yet crossed the horizon. Because your time has been running so slowly near
the hole, you’ll return to find me much older than you or even long gone as
you come back to your starting place centuries, millennia, or even farther in
the future. But for you only a matter of hours or days might have elapsed. The
numerical details depend on the mass of the hole and how close you let
yourself get. In any event, here’s another way—like the twins example of
special relativity—for you to leapfrog your way into the future. Again,
there’s no going back if you don’t like what you find!
So far I’ve described your black-hole journey largely from my viewpoint
as an observer distant from the hole. But what’s it like for you? Assuming
you don’t decide to return, you’ll find yourself descending toward the hole
and right through the event horizon in a finite and possibly even very short
time. Will you experience your clock slowing, your movements becoming
languid, and your body remaining forever youthful? Not at all! That would
violate the Principle of Relativity. Your freely falling or free-float reference
frame is a perfectly good one for doing physics, and your clock, your body,
and all other physical processes should seem perfectly normal to you. Will
something dramatic happen at the moment you cross the event horizon? No!
Again, the relativity principle asserts that the laws of physics work just
perfectly in your free-float reference frame, so if something odd happened
right at the horizon, that would violate relativity. You wouldn’t even know
when you had fallen past the point of no return.
There is, however, a practical constraint on your imaginary black-hole
journey. For a black hole formed from a collapsed star (more on this in the
next section), the curvature of spacetime would become so great that your
body would be torn to pieces before you crossed the horizon. Figure 14.3
showed how the spacetime curvature that constitutes Einstein’s gravity
results from what Newton would describe as differences in Newtonian
gravity from place to place. As Figure 14.3 suggests, the consequence in
either Einstein’s or Newton’s picture is a force that stretches a freely falling
body in the vertical direction and compresses it horizontally. For a star-mass
black hole, those forces become fatally large well outside the event horizon.
For black holes of a billion star masses, though—holes now believed to
inhabit the centers of some galaxies—the tidal forces on a human body would
be negligible at the event horizon. You would fall freely across the horizon of
such a hole with no sense whatsoever of anything unusual.
Once inside any black hole, however, you’d be drawn inexorably toward
the singular point at the hole’s center. No force in the Universe is strong
enough to resist the pull or, in general relativistic terms, to prevent your
worldline through spacetime from intersecting the singular point. Sooner or
later (sooner for a lower-mass hole, later for a larger one) spacetime
curvature will become significant over the size of your body and you’ll be
destroyed.

Do Black Holes Exist?

Black holes remained figments of theorists’ imaginations until the late

twentieth century. Then, with the advent of space-based astronomy in the
1960s, astrophysicists suddenly had new windows on the Universe.
Particularly significant in the discovery of black holes was the x-ray
window. X-rays, similar to those used in medical imaging, don’t penetrate
Earth’s atmosphere. When astrophysicists first turned satellite-borne x-ray
detectors on the cosmos, they were surprised to find a number of bright,
pointlike x-ray sources. Many of these proved to be binary stars in which x-
rays were produced in the intense heat generated as gas flowed from a large,
visible star to an unseen companion. Those unseen companions became the
first candidates for black holes.
Why should a black hole become a cosmic x-ray source? Because black
holes, and their neutron-star cousins, are so deep gravitationally that infalling
matter accelerates to enormous speeds. If that matter is gaseous, friction in
the flowing gas generates very high temperatures—so high that instead of
glowing red-hot like a stove burner, or white-hot like our Sun, the infalling
gas “glows” in x-ray “light.” (Recall from Chapter 4 that x-rays are just
another form of electromagnetic waves, distinguished from visible light by
their much shorter wavelength.) Some x-ray sources showed rapid periodic
fluctuations, indicating that the unseen companion was a rapidly spinning
neutron star—a pulsar, as introduced earlier in this chapter. Others showed
no such pulses. Furthermore, astrophysicists can compute the masses of stars
in a binary system, and in some systems the mass of the companion star
exceeds theorists’ upper limits for the mass of a neutron star. Here, black
holes provide the simplest and most coherent explanation of the
observations.
Evidence for black holes in binary systems grew stronger through the late
twentieth century, and today most astrophysicists acknowledge the existence
of black-hole binaries. Although these systems are too distant for us to see
their details directly, decades of increasingly finer observations and
theoretical modeling have given us a firm picture of what a black-hole binary
must look like. As Figure 15.6 shows, the salient features are a large, visible
star and an invisible black hole in close orbits around each other. Star and
hole are so close that gas flows from the former to the latter. Because of the
system’s orbital motion, the gas doesn’t flow straight but forms a swirling,
donut-shaped cloud around the black hole. This accretion disk is where the
intense heat and x-rays are generated. The gas loses energy through the heat-
generating friction, allowing it to spiral ever closer to the black hole.

Fig. 15.6 What a black hole in a binary sy stem might look like. The massivestar in the center is distorted by the strong tidal
forces of the black hole. Gas from the star is drawn to the vicinity of the hole, where it swirls around in a disk before finally
disap p earing into the hole itself. Friction in the gas generates such high temp eratures that the gasp roduces cop ious x-ray
emission.

Even as some astrophysicists studied the newly found binary x-ray

sources in our own galaxy, others sought to understand some of the most
puzzling and distant objects ever found—the quasars. These enigmas first
appeared as pointlike images at the very edge of the visible Universe, yet
they emit more energy than an entire galaxy comprising hundreds of billions
of stars. What could possibly power such colossal energy sources?
 Higher resolution images soon showed that quasars’ energy generation
regions must be very small, and again astrophysicists’ imaginations turned to
black holes. Might the accretion of matter into the hole account for a quasar’s
extreme energy output? Another clue came from images showing some
quasars emitting high-speed jets of material. A newly discovered solution for
the structure of spacetime around a rotating black hole suggested that such
jets would be a natural feature of a rotating hole system, with matter
accreting near the “equator” of the spinning hole and jets emerging at its
“poles.” Furthermore, quasars seemed similar to so-called active galaxies,
which also exhibit jets and substantial energy output from their galactic
centers. Could the same black-hole mechanism be at the heart of both sorts of
objects?
With the advent of the Hubble Space Telescope and high-tech Earth-
based telescopes late in the twentieth century, astronomers observing nearby
galaxies found similar compact, high-energy sources at the centers of many
galaxies, including our own. The high speeds of stars and gases swirling
around these centers again suggest black holes as the ultimate energy source.
By now, it appears that quasars, active galaxies, and normal galaxies like our
Milky Way probably all contain massive central black holes. The quasars are
believed to be the cores of galaxies so distant that we’re seeing some as they
were billions of years ago, at a time when accretion into their central black
holes was particularly active and their energy output exceptionally strong.
Active galaxies have less vigorous accretion rates, and in normal galaxies
the process is quite modest.
How big are the black holes at galactic centers? The hole in our Milky
Way contains several million Suns’ worth of matter. Holes lurking in some
quasars may exceed a billion solar masses. In physical dimension, these
holes would be about 10 times the diameter of our Sun and roughly the size
of the Solar System, respectively. Should you decide to explore these holes
by dropping freely into them from a great distance, you would have about 10
seconds between crossing the horizon and reaching the singularity of the
Milky Way’s hole and a couple of hours for the quasar’s hole. Of course
these black holes continue to grow, on astronomical timescales, by gobbling
up stars and gas from the high-density surroundings of their galactic-center
neighborhoods.
Ripples in Spacetime
Drop a rock into a pond and circular ripples spread across the water. Only
later, when the ripples have reached it, can a distant point on the pond
“know” about the rock’s plunge. Disturb spacetime, perhaps through a violent
collision between black holes, and what happens? Special relativity assures
us that distant points can’t know instantaneously about this event; indeed,
that’s one of the reasons Einstein knew that Newton’s theory of gravity
couldn’t be right. As on the pond, “ripples” in spacetime itself propagate
outward, carrying information about the violent event at their center. What’s a
ripple in spacetime? Simply a change in the curvature of spacetime—a
change that moves outward from the disturbance that initiates it. Einstein’s
general relativity predicts the existence of these ripples in spacetime; they’re
called gravitational waves. General relativity also shows that gravitational
waves travel at the familiar speed of light, c.
Gravitational waves are an entirely new phenomenon predicted by
general relativity. Detecting these waves would provide an independent
confirmation of Einstein’s theory, and might give us a novel window on the
cosmos. How can we detect them? With their peaks and troughs, gravitational
waves would stretch and compress spacetime itself. We could detect the
spatial part of that stretch and compression by measuring the associated
motion of physical objects. Because mass is what responds to gravity—that
is, to spacetime curvature—we’ll have better luck with massive objects. For
some decades researchers around the world have attempted to detect
gravitational waves by using huge aluminum bars equipped with exquisitely
sensitive motion detectors. The bars would be set into vibration by a passing
gravitational wave. To eliminate vibrations induced by trucks, scientists
walking by, and other mundane causes, a typical experiment involves
identical setups located thousands of miles apart. The only events considered
real candidates for gravitational waves are those that trigger both detectors.
Although these experiments have produced a few intriguing signals, none to
date has passed muster as a true gravitational-wave detection.
This may all change soon, however, as a new generation of gravitational-
wave detectors becomes operational. Abandoning the massive cylinders of
first-generation detectors, gravitational-wave researchers are now turning to
interferometry—the method pioneered by Michelson in his famous
experiment with Morley—to measure precisely the distance between two
widely separated objects. Recall from Chapter 6 that a slight change in the
travel time for light along one arm of the Michelson–Morley apparatus
(shown in Figure 6.2) would result in a shift in the observed interference
pattern. Michelson and Morley hoped to find changes associated with
differences in the speed of light; for gravitational-wave detection, we’re
looking for changes in the distance from beam splitter to mirror as a
spacetime ripple goes by.
In the United States, the Laser Interferometer Gravitational-Wave
Observatory (LIGO) consists of two complete interferometers, one in
Washington state and one in Louisiana. Each consists of a Michelson-type
apparatus with two perpendicular arms 2.5 miles long. These instruments can
measure changes in the lengths of these 2.5-mile paths of less than a trillionth
the diameter of a human hair, and with that sensitivity they should be able to
detect gravitational waves produced in the supernova explosions that result
in neutron stars and black holes; in collisions of black holes and neutron
stars; and in the Big Bang explosion that began our Universe. Similar
detectors are being built around the world, and collectively they will give
astrophysicists a new type of “telescope” for observing hitherto unseen
events in the cosmos.
Even more ambitious is the proposed Laser Interferometer Space
Antenna (LISA), a Michelson-type apparatus whose “arms” will consist of
spacecraft forming a triangle 3 million miles on a side (that’s 400 times
Earth’s diameter, or 3 percent of the Earth–Sun distance!). Sensitive to
changes in that distance on the order of a billionth of an inch, LISA should
“see” gravitational waves from beyond our galaxy, including those generated
by the massive black holes at the centers of other galaxies.
Although gravitational waves have yet to be detected directly,
astrophysicists nevertheless have one piece of convincing evidence for their
existence. This is the Taylor–Hulse binary pulsar, which I introduced early in
the chapter for its obvious general-relativistic orbit precession. Recall that
the orbital period of this binary neutron star admits very precise
measurement and that this period is changing slowly with time. Why
changing? Because the neutron stars’ orbits are shrinking. Why shrinking?
Because the stars are losing energy, gradually spiraling closer just as a
satellite in orbit near Earth slowly loses altitude through friction with the
upper atmosphere. But there’s no atmospheric friction in the binary pulsar.
Instead, the neutron stars lose energy through an unseen process that leaves
its fingerprint in the slowly decaying orbital motion. That process is the
generation of gravitational waves. As the massive neutron stars swing round
in their close orbit, the energy they expend in disturbing spacetime—energy
that’s carried away as the yet-undetected ripples of gravitational waves—is
lost from their orbital motion. Years of meticulous observation show that the
binary pulsar is losing energy at just the rate that general relativity predicts.
So although we haven’t “seen” gravitational waves from the binary pulsar,
we’re quite sure they’re being produced and are responsible for the observed
orbital changes. Incidentally, their meticulous observations of the binary
pulsar and its general relativistic implications earned Taylor and Hulse the
1993 Nobel Prize in Physics.

Relativity in the Astrophysicist’s Toolbox

The observational technologies and physical theories behind modern
astrophysics reveal a Universe rich in strange and wondrous phenomena,
some so strange that they would have seemed unimaginable only a few
decades ago. Many of these phenomena—from black holes to binary pulsars,
from gravitational waves, lenses, and time dilation to unusual orbits in
strongly warped spacetime—are solidly confirmed and regularly observed.
Others, like the wormholes of Carl Sagan’s Contact, remain speculative.
They have some grounding in theory, though no supporting observational
evidence. But all these phenomena share common roots in the extraordinary,
flexible spacetime geometry unveiled by Einsten’s relativity. For that reason,
relativity has become a working tool for many of today’s astrophysicists.
Even those who don’t observe or theorize about relativistic objects may still
find relativistic tools useful, as when gravitational lenses aid in the search
for Earthlike planets circling distant stars. Relativity remains the ultimate
tool for those who seek to understand the really Big Picture—the workings of
the entire Universe. We’ll go there in the final chapter.
CHAPTER 16

EINSTEIN’S UNIVERSE
• • •

The Big Picture

Gravity isn’t just about planetary orbits or even exotic black holes. Gravity
binds stars into galaxies, galaxies into clusters of galaxies, and clusters into
superclusters. Structure persists to the largest scales, with “walls” and
“bridges” where galactic superclusters concentrate. Between these are
sparsely populated voids. Had we evolved in a galaxy in such a void, it
would not have been until the end of the twentieth century that our
astronomical technology would have been capable of detecting even the
nearest neighbor galaxies. But then, what incentive would we have had to
develop that technology?
How did structure evolve in a Universe that, at its beginning, was a
simple, homogeneous soup of matter and energy? What is the shape of the
Universe today? What is its ultimate fate? On the vast scales of distance and
mass that describe the Big Picture configuration of today’s Universe, gravity
is the dominant influence. Our description of gravity—the general theory of
relativity—is therefore at the heart of cosmology, the study of the overall
structure of the Universe.

Einstein’s Blunder
Einstein himself was among the first to apply general relativity to cosmology,
and what he found was unsettling. The Universe, according to the simplest
formulation of general relativity, couldn’t be static; it had to be expanding or
contracting. But prevailing wisdom held that the Universe was static, having
existed forever unchanged in its overall features. Now, in the mathematical
development of general relativity, there arose a number called the
cosmological constant, whose value seemed to be arbitrary. Absent any
reason to the contrary, the most sensible choice is to set this number to zero,
giving the simplest formulation of the theory. That choice is what made
Einstein’s universe expand or contract, so he introduced a nonzero
cosmological constant of just the right value to keep the Universe static.
Einstein’s cosmological constant represented a sort of repulsive force acting
on the largest scales, preventing the Universe from collapsing under its own
gravity. Nothing then known about gravity suggested such a repulsion, so
Einstein was a little uneasy with his cosmological constant. But it was not
inconsistent with the general theory, so he accepted the cosmological
constant as necessary to make his theory fit what seemed to be the real
Universe.
Other physicists also explored relativity-based models for the Universe.
In the early 1920s, the Russian Aleksandr Aleksandrovich Friedmann found
solutions to the equations of general relativity that described an evolving
universe beginning in a very dense state and then expanding at an ever
slowing rate. Friedmann’s results revealed two distinct kinds of possible
universes: those in which expansion continues forever, albeit at an ever
slower rate; and those in which the expansion eventually halts and the
universe subsequently contracts to an eventual high-density crunch. These
two possible fates are intimately linked to the overall spacetime geometry of
the Universe. A forever expanding Friedmann universe has negative
curvature, meaning its spacetime is shaped like the four-dimensional analog
of a saddle or a pass between mountain peaks, and is infinite in extent (see
Figure 16.1a). A universe that eventually collapses has positive curvature,
like the surface of a sphere (Figure 16.1b). It’s closed back on itself and is
finite in extent. Yet, like a sphere’s surface, it has no edge. Between these
two types of Friedmann universe is a dividing case. Overall, it’s flat
(although spacetime in such a universe would still be curved locally in the
vicinity of massive objects), infinite, and will just barely expand forever.
Fig. 16.1 (a) A universe with negative curvature would be analogous to thissaddleshap ed surface in three dimensions. (b) A
p ositively curved universe would be analogous to a three-dimensional sp here.

These geometrical analogies are imperfect. You can imagine a sphere or

saddle shape by picturing it in three dimensions, and I’m asking you to think
of the four-dimensional spacetime Universe as being analogous to the surface
of the sphere or saddle. But that “surface” is all there is; there’s no fifth
dimension for it to be curved in. And the four-dimensional Universe may be
expanding. Into what, you ask? No answer; it’s the fabric of space and time
itself that’s expanding.
By the late 1920s, the American astronomer Edwin Hubble had
completed a study of nearby galaxies and found that they were receding from
us at speeds directly proportional to their distances. Hubble concluded that
our Universe is indeed expanding. (If this sounds like we’re at the center—a
disturbingly un-Copernican thought—rest assured. In an infinite Universe
there’s no center, and if ours is a finite Friedmann universe there’s still no
center—just as there’s no point on the surface of the Earth that can claim to
be the center.) So the cosmological constant wasn’t necessary after all, and in
the 1930s Einstein abandoned the constant as “the greatest blunder of [his]
life.”
Through the rest of the twentieth century, astrophysicists accumulated
ever more evidence, both observational and theoretical, for a Universe that
began in a Big Bang explosion some 14 billion years ago and has been
expanding ever since. Much of this evidence lies outside the realm of
relativity, and I’ll leave it to others (see the Further Readings list) to present
that story. Yet, at the heart of modern cosmology lies general relativity,
because it’s the interaction between matter and spacetime that shapes the
overall geometry of the Universe.
The Fate of the Universe
One big question in cosmology is whether the Universe will expand forever
or will ultimately contract. Absent a cosmological constant, general relativity
shows that this question is equivalent to another: What is the large-scale
curvature of spacetime? If it’s positive, like the surface of a sphere, then the
Universe is closed and will ultimately contract to a Big Crunch. If it’s
negative, like a saddle-shaped surface, then the Universe is open and will
expand forever. And if it’s flat—meaning the rules of tenth-grade Euclidean
geometry apply on the large scale—then the Universe will just barely expand
forever.
What determines the geometry and thus the ultimate fate of the Universe?
Because matter is what warps spacetime, giving it curvature, the answer lies
in the amount of matter present in the Universe. There’s a critical density,
averaging on the order of a few hydrogen atoms in each cubic meter of
otherwise empty space, above which the amount of matter is enough to curve
spacetime back on itself, giving a closed, finite, positively curved Universe
that will eventually collapse. If its density is below critical, the Universe
will be open, negatively curved, and will expand forever. Right at the critical
density, the overall geometry of the Universe is flat and it will just barely
expand forever.
So which is it? Remarkably, the real Universe seems close enough to the
dividing case of critical density that the answer isn’t obvious from
observations. The visible matter—stars, interstellar gas clouds, galactic
centers, and everything else that emits electromagnetic radiation—constitutes
only a small fraction of the matter needed to achieve critical density.
However, the movement of stars within galaxies suggests that unseen dark
matter also contributes to the Universe’s overall density. What is that dark
matter? Some of it could be mundane things like burned-out stars too faint to
see, or isolated black holes, but theory suggests that most of it must be in the
form of undiscovered exotic matter different from the ordinary matter that we
see throughout the Universe. Here’s a humbling thought: perhaps most of the
Universe is made of something we know essentially nothing about!
Today, most cosmologists believe there’s enough dark matter that the
Universe has exactly the critical density and that its overall geometry is flat.
A number of theoretical arguments suggest a flat Universe, and recent studies
of the so-called cosmic microwave background radiation—a kind of fossil
remnant of times early in the Big Bang—support the flat-Universe hypothesis.
That would end the Big Picture story of spacetime geometry, except for a
remarkable discovery made in 1998. Astrophysicists had been studying
supernova explosions in very distant galaxies, hoping to use their measured
brightness to infer the rate at which the expansion of the Universe is slowing.
That method works because distant objects appear from Earth as they were a
long time ago, as long ago as it took light from them to reach us. Comparing
the brightness and recession speed of supernovas at different distances thus
gives the expansion rate of the Universe at different times. To the
astrophysicists’ great surprise, they found that the expansion wasn’t slowing
at all; rather, it was accelerating! It appears that this acceleration began about
5 billion years ago, coincidentally around the time the Sun, Earth, and the rest
of our Solar System formed.
What could cause an acceleration of the cosmic expansion, despite the
slowing effect of mutual gravitation? One possible answer goes right back to
Einstein: His cosmological constant provides a cosmic repulsion that could
increase the expansion rate. That’s only one possible answer, and
cosmologists are proposing new phenomena with names like “quintessence”
and “dark energy” that could also produce an accelerated cosmic expansion.
With more detailed observations of ever more distant realms becoming
accessible to our telescopes, and after much theoretical analysis, we’ll
eventually learn whether Einstein was right after all when he introduced the
cosmological constant into general relativity.

A Theory of Everything?
General relativity provides our best understanding of the Big Picture of the
Universe as a whole, but can general relativity tell us everything? The
answer is no. That’s because physicists have not yet been able to reconcile
general relativity with quantum physics, the theory that describes matter on
atomic and smaller scales. Special relativity and quantum physics were
reconciled decades ago, giving us a powerfully accurate description of the
behavior of matter at small scales that reveals some entirely new atomic
phenomena required by the Principle of Relativity. A reconciliation of
general relativity and quantum physics, though, faces deep conceptual
problems. That’s because the essence of quantum physics is quantization,
meaning that the “stuff” of the Universe—from particles of matter to energy
itself—comes in discrete “chunks” rather than being continuously
subdividable. You can have one electron, but not half of one. You can have a
“chunk” of light energy of a given color (called a photon), but you can’t have
less. It’s this essential graininess that ultimately dictates the strange rules
governing the quantum world.
We, and the things we interact with in everyday life, are so large that we
don’t notice quantization. A glass of water contains so many individual water
molecules that the quantization of the water into molecules doesn’t seem to
make a difference; the water might just as well be a continuous fluid. A light
bulb or even a candle flame emits so many photons that they might as well
constitute a continuous stream of energy. All this is even more true for the
planets, stars, and galaxies that make up the astrophysical Universe. As a
practical matter, astrophysics’ description of the Universe is at such a large
scale that the reconciliation of quantum physics and general relativity is
usually unimportant. Put another way, the curvature of spacetime is generally
significant only on scales vastly larger than the size of atoms or elementary
particles. But we can imagine situations where this is not true, situations
where spacetime is so tightly curved that even something as small as an
elementary particle is big enough to experience spacetime curvature. What
absurd situations would those be? One is the singularity at the center of a
black hole. There, general relativity predicts that spacetime curvature
becomes infinitely sharp. Before that true singularity is reached, effects of
quantum physics must come into play. Another example is the very early
Universe, at the start of the Big Bang—specifically, the time before about
10–43 of a second from the beginning. (That’s 1/10 000 000 000 000 000
000 000 000 000 000 000 000 000 000 of a second, a time known as the
Planck time.) At that point the Universe was so dense that even its
geometrical structure would have to be described using quantum physics.
Because we don’t know how to reconcile quantum physics and general
relativity, we can’t say much about conditions at the center of a black hole or
in the early Universe before the Planck time. Ultimately, our knowledge is
incomplete.
 Why are the two pillars of modern physics—general relativity and
quantum physics—so seemingly irreconcilable? Because general relativity
is, at its heart, a continuous theory. It envisions a spacetime that is
continuously divisible. That means we can divide a meter or an inch into
ever smaller lengths, without limit. Similarly, we can divide the 1-second
interval between ticks of a clock into as many ever tinier sub-intervals as we
wish. For time intervals much longer than the incredibly tiny Planck time
introduced in the previous paragraph and for distances much longer than the
Planck length—or the distance that light travels in one Planck time (about
10–35 of a meter)—the idea of a continuous spacetime is a very good
approximation to reality. As we approach the Planck length and time, though,
spacetime itself must exhibit quantization. There must be fundamental,
indivisible units of length and time on the order of the Planck units. What
does this mean? No one knows for sure. Because the quantum world is
seething with submicroscopic events, physicists often picture quantized
spacetime as a spongelike structure with a complicated and everchanging
geometry, in which wormholes and bridges continually form and dissolve.
Will we ever reconcile general relativity and quantum physics, producing
a successful theory of quantum gravity? Probably. Some physicists believe
we’re close, with a group of theories collectively called string theory or M-
theory, which envision the fundamental entities of nature not as particles but
as tiny looplike strings. Vibrations of these strings represent the different
elementary particles of physics. Because the strings have finite size, even
though they’re indivisible, they manage to sidestep problems with gravity at
the Planck scales. Furthermore, some versions of the theory suggest particles
called gravitons that would quantize gravity in the same way the photon is a
quantized chunk of electromagnetic energy. No string theory yet conforms to
the observed set of elementary particles and all require a remarkable
complication—the inclusion of at least six more dimensions in addition to the
familiar three dimensions of space and one of time. Furthermore, no one has
yet figured out how to test experimentally most predictions of string theories.
For some physicists, string research is nevertheless getting us close to a
“Theory of Everything,” explaining the behavior of the entire Universe from
the smallest to the largest scales. To others, string research seems a fruitless
mathematical exercise. Time will tell.
Out of the Black Hole: Inflation, Multiple Dimensions, and
Parallel Universes

Although no one has yet figured out how to merge general relativity and
quantum physics into a consistent theory, we do have some hints of how the
quantum realm might influence relativity. The British physicist Stephen
Hawking—occupant of Newton’s former chair at Cambridge University—has
shown how quantum effects lead to a subtle “glow” of radiation from the
spacetime around a black hole. That radiation ultimately saps the black hole
of its energy—equivalently, by E = mc2, of its mass—causing the hole to
“evaporate” and eventually disappear altogether. For the massive black holes
that astrophysicists have discovered, that process will take far longer than
the present age of the Universe. However, tiny black holes that might have
formed early in the Big Bang would already have vanished through this
mechanism of Hawking radiation.
Quantum physics shows that the emptiness we call vacuum isn’t quite
empty, but is seething with ghostly particle–antiparticle pairs that burst into a
fleeting existence and promptly annihilate. Hawking radiation arises when
one of these particles falls into a nearby black hole, leaving its partner no
one to annihilate with. The lonely partner then looks like a particle that’s
emerged from the hole. Although that isn’t quite the case, the black hole has,
in a sense, created the new particle and thus has given up some of its own
mass-energy.
Not only is the seemingly empty vacuum actually the site of frenetic
physical activity, but the current theory of elementary particle physics—the
so-called standard model—also suggests the possibility of a “false vacuum,”
a high-energy state that might decay rapidly to a true vacuum. Decay of the
false vacuum can lead to exponentially rapid inflation of spacetime. Detailed
theories of the Big Bang suggest that a period of just such inflation occurred
in the first fraction of a second of the Universe. During this brief instant, the
Universe grew in size by many orders of magnitude—a process that flattened
out any curvature and resulted in a Universe whose large-scale spacetime
geometry is essentially the flat, Euclidean geometry that you learned in tenth
grade.
 It’s the prospect of inflation that leads to a remarkable possibility I
mentioned in Chapter 1: a baby universe could bud from a parent and then
grow by inflation to become a full-blown universe in its own right. Repeated
endlessly through eternity, this process would give rise to a complex
Multiverse of which our Universe is but one branch.
Attempts to merge relativity with quantum physics often seem to lead us
beyond the three familiar dimensions of space and one of time. Today’s
eleven-dimensional string theories are only the latest instance of
multidimensional theories of everything. Einstein himself spent much of his
later life working, in vain, on five-dimensional theories that he hoped would
combine his general relativity with quantum physics.
Extra dimensions allow for the remarkable possibility that there may be
parallel universes lying very close to ours but separated by an extra
dimension that we don’t directly perceive. Stanford physicist Savas
Dimoupolos and colleagues have argued that a parallel universe might be
only a few millimeters (about a tenth of an inch) away from ours (Figure
16.2). We think we know nothing of it because all processes in that universe
are confined to its three dimensions of space and one of time. But not quite!
According to Dimoupolos, the quantized particles of gravitational influence
—the gravitons—would be unique in that they could traverse the extra
dimension. So, while ordinary matter, light, and other forms of energy would
be confined to that nearby but hidden universe, its gravitational influence, as
Figure 16.2 shows, would not. If this idea is correct, then what we infer to be
dark matter in our Universe may actually be ordinary matter in the parallel
universe! We perceive only its gravitational influence, not the matter itself,
and so of course it seems dark to us. Strange as this parallel-universe idea is,
no one has yet disproved it, and experiments with gravity over millimeter
scales might just reveal that extra dimension.
Fig. 16.2 Parallel universes? Each p lane is a two-dimensional analog of a universe with three sp ace dimensions, p op ulated with
galaxies. The p lanes shown are millions oflighty ears in extent, but they ’re sep arated by a fraction of an inch in an extra
dimension that we don’t directly p erceive. Thin arrows rep resent electromagnetic interactions(e.g., light beams) that are
confined to the individual universes. Gravity (thick arrows) extends into the extra dimension, so each universe feels the other’s
gravitational effect. (Galaxy p hotos courtesy of NASA.)

Because spacetime can be curved, it’s also possible that the parallel
universe isn’t really a distinct universe but rather a part of ours that’s very
far away in the ordinary dimensions but close in the extra dimension. If that’s
true, then it’s possible to imagine a short wormhole connecting our
neighborhood to the distant realm of the parallel universe. Maybe that’s what
the machine in Carl Sagan’s Contact is all about.

Intuiting Relativity

We’ve now come full circle, back to some of the stranger ideas I introduced
briefly in the first chapter. Those ideas, and other new concepts ranging from
the relativity of simultaneity to black holes and gravitational waves, become
possible when we abandon rigid, absolute space and time and replace them
with the flexible, curving spacetime of Einstein’s relativity. Einstein’s
special and general theories of relativity reveal a Universe far richer and
stranger than anything in our commonsense experience. In this book I’ve
taken you rather thoroughly through the special theory, trying to convince you
of its validity on rigorous logical grounds. Yet I’ve continually reminded you
that you’ll never have a fully natural intuition for relativity. I don’t either, and
I don’t believe Einstein did. That’s because none of us has experienced in
everyday life the conditions that show up the difference between Einsteinian
relativity and Newtonian common sense.
With general relativity I’ve had to leave things a lot murkier. I’ve given
you the briefest motivation for the fundamental underpinnings of the theory,
ultimately leading to the idea of gravity as the geometry of spacetime. In
these final chapters I’ve described some of the consequences of general
relativity—especially significant where gravity is strong—but I haven’t at
this level been able to give a lot of motivation for those consequences, let
alone any kind of intuitive feel for them.
So you can understand relativity logically, and, if you work with it
enough, gain a comfortable familiarity. For the nonscientist that’s possible
with the simpler concepts and mathematics of special relativity, and for the
mathematically brave a thorough understanding of general relativity is
possible too. But in neither case will you have much natural, commonsense
intuition about relativity. Is such intuition at all possible? I believe it would
be if we grew up experiencing the Universe in the full richness that relativity
describes. If, as a baby, you crawled at speeds approaching that of light, then
your common sense would be fully consistent with special relativity. You
would have no misleading notions about absoluteness of space and time
measurements, and reference frame–dependent lengths, times, and even
simultaneity would be the norm for you. You wouldn’t be surprised when a
friend went out for a high-speed jog and came back a member of a younger
generation. In short, special relativity would make perfect sense to you
because it would be part of your regular experience.
If, in addition to being a relativistic crawling baby, you were so large
that your body directly experienced the curvature of spacetime, then the
geometrical nature of gravity would also be intuitively obvious to you. Your
geometry teacher, drawing on a blackboard so big that spacetime curvature
was obvious, would never deceive you with Euclidean nonsense; it would be
obvious that the angles of a triangle add to other than 180 degrees and that
parallel lines might meet (or might not, depending on the sign of the local
spacetime curvature). You would experience directly that these geometrical
effects arose in the presence of matter and energy, and your understanding of
gravity would naturally be that of Einstein. But you’re small in relation to the
spacetime curvature in your neighborhood and you move slowly relative to
things important in your life; thus, for you, relativity can never seem intuitive.
You can, however, grasp the simple principle at the heart of all relativity—
that motion doesn’t matter, or that there’s no preferred frame of reference—
and despite the failure of your intuition, you can understand intellectually the
remarkable consequences of this principle and appreciate the wonderfully
rich Universe it engenders.
APPENDIX

TIME DILATION
• • •

Here I’m going to use high-school math to convince you that the formulas for
time dilation and length contraction do indeed follow directly from the
Principle of Relativity. You can take those formulas on faith and skip this
appendix, but if you’re a stickler for logical consistency, then reading this
will give you what you need.

High-School Math Review

At some point in your school career, either in junior high or early high
school, you learned the Pythagorean theorem, which gives the length of the
hypotenuse (long side) of a right triangle in terms of the lengths of the other
two sides. Figure A.1 shows a right triangle with sides labeled A, B, and
hypotenuse C. The Pythagorean theorem says

C2 = A2 + B2.

You also learned that you can do the same thing to both sides of an equation
and it’s still a valid equation. So let’s take the square root of both sides, to
get the length of the hypotenuse, C:

C = √A2 + B2.

I’ve marked this value on Figure A.1.

Fig. A.1 The Py thagorean theorem gives the hy p otenuse of a right triangle in terms of the lengths of the other two sides.

You also learned, probably even before high school, that distance =
speed × time. After all, that’s just what speed means: go 50 miles per hour
for 2 hours, and you’ve gone 100 miles. Distance equals speed times time.

Time Dilation

Now you’re all equipped to understand time dilation, quantitatively. Figure

A.2 is a modified version of Figure 8.3, the light box that I used to introduce
time dilation. I’ve emphasized the light path by making it a solid line and
dimming everything else. In Figure A.2a the box is again shown in a
reference frame where it’s at rest, and it’s clear that the light goes a round-
trip distance 2L between source, bouncing off the mirror, and returning to the
source. We’ve called the time in this reference frame t’ (“t prime”), and the
light’s speed is, of course, the speed of light, c. So the formula distance =
speed × time becomes

2L = ct’.
Fig. A.2 The light box of Figure 8.3, shown with various quantities used to derive the time-dilation formula

(Here I’m adopting the convention of implied multiplication, where two

symbols written next to each other are assumed to be multiplied. This avoids
use of the multiplication sign.) Dividing both sides of this equation by c gives

That was easy!

Now let’s go to the frame in which the box is moving (Figure A.2b). The
box is moving at speed v, and we’re calling t the time between departure and
return of the light flash. So distance = speed × time tells us that the distance
the box travels is just vt. While the box goes half this distance, the light itself
takes the diagonal path up to the mirror, as shown on Figure A.2b. How long
is this diagonal? It’s the hypotenuse of a right triangle with sides ½vt and L.
The Pythagorean theorem then gives its length:

√(1/2vt)2 + L2, or √1/4v2t2 + L2 .

The total light path is twice this distance, the time the light takes is t, and—
here’s where relativity comes in—the speed of the light is c in this reference
frame as well. So distance = speed × time now gives

2 √1/4v2t2 + L2 = ct.
Now, we want to know the time t. Unfortunately it’s on both sides of this
equation and tangled up in a square root. To get it out, square both sides:

4(1/4v2t2 + L2)= c2t2, or v2t2 + 4L2 = c2t2.

Now subtract v2t2 from both sides:

4L2 = c2t2 − v2t2, or 4L2 = c2t2 (1 − v2/c2).

Almost done! Divide both sides by c2 and take the square root of both sides.
The result is

But 2L/c is just the time, t’, measured in the frame at rest with respect to the
box [see Equation (A2)]. So we have

t’ = t√1 − v2/c2.

This is just the time-dilation formula I presented in Chapter 8, except that it

has c in it. In Chapter 8 I defined the speed v as a fraction of c, so the
quantity v/c here is the v of Chapter 8. Thus we’ve reached Chapter 8’s
formula, t’ = t√1 − v2 and we’ve done so in a way that follows rigorously
from the Principle of Relativity.
GLOSSARY
• • •

Aberration of starlight The change in ap p arent p osition of a star, due to Earth’s orbital motion. Used to show that Earth did
not drag ether with it.
Acceleration The rate at which an object’s motion changes. Acceleration includes changes in sp eed or direction.
Accretion disk The disk-shap ed cloud of matter swirling around and into a black hole.
Action at a distance The Newtonian view that influences, esp ecially gravity, reach instantaneously from a gravitating object
to more distant objects. The action-at-a-distance p icture is inconsistent with sp ecial relativity.
Antimatter M atter consisting of p articles with p rop erties that are exactly the op p osite of ordinary matter. Antielectrons
(p ositrons), for examp le, are like electrons but with p ositive charge. Antimatter is created in energetic interactions involving
elementary p articles or high-energy electromagnetic radiation. When they meet, matter and antimatter annihilate in a burst
of energy.

Beam splitter A device that sp lits a beam of light into two beams traveling on different p aths. The simp lest beam sp litter is a
mirror with insufficient reflective material.
Big Bang The cosmic exp losion that began the Universe.
Big Crunch In theories that p rop ose an oscillating Universe that exp ands and then contracts, the Big Crunch is the ultimate
state of contraction before the next exp ansion. In the Big Crunch, all matter would be comp ressed to near-infinite density.
Black hole An object so massive y et so comp act that not even light can escap e its gravity.

Cosmological constant A number Einstein introduced into the general theory of relativity so the theory would p redict a static
Universe. Einstein abandoned the cosmological constant when observations showed that the Universe was in fact
exp anding. Discoveries at the end of the twentieth century suggest the constant might be necessary after all.
Cosmology Study of the origin, evolution, and large-scale structure of the Universe.
Critical density The minimum average density necessary for the Universe to exp and forever instead of eventually collap sing.

Dark matter Unseen material believed to make up over 90 p ercent of the Universe’s mass.
Doppler effect The increase (or decrease) in the frequency of waves—sound or light—when the source of the waves is moving
toward (or away from) the observer.

Electric charge A fundamental p rop erty of matter that is at the basis of all electric and magnetic interactions.
Electromagnetic induction The p henomenon whereby a changing magnetic field p roduces an electric field.
Electromagnetic wave An electromagnetic p henomenon in which changing electric and magnetic fields continually generate
each other, p roducing a wave of electromagnetism that travels with the sp eed of light, c. Electromagnetic waves include
radio waves, microwaves, infrared, visible light, ultraviolet, x-ray s, and gamma ray s.
Electron A fundamental p article in nature. Carries a negative electric charge and little mass.
Elsewhere The elsewhere of an event consists of those events that cannot influence or be influenced by the given event.
Energy One of the two manifestations of mass-energy, the fundamental “stuff” of the Universe. Energy takes many forms,
including the energy of motion, gravitational energy, heat energy, electromagnetic energy, etc. Without energy there would
be no motion, no activity whatsoever.
Epicycle Smaller circular p aths of the p lanets, imp osed on their larger circular orbits and required in early cosmological models
to account for retrograde motion.
Escape speed The sp eed necessary for an object to escap e to an infinitely great distance from a gravitating body. At Earth’s
surface, escap e sp eed is about 7 miles p er second; at the horizon of a black hole, it’s the sp eed of light.
Ether Hy p othetical substance p rop osed by nineteenth-century p hy sicists as the medium of which light waves are a
disturbance.
Event Something that hap p ens, characterized by where it occurs in sp ace and when it occurs in time.
Event horizon The region around a black hole at which escap e sp eed becomes the sp eed of light. Nothing—no material object,
no light, and no information—can escap e from within the event horizon.

Field An invisible influence in the sp ace surrounding a massive object (gravitational field), an electrically charged object
(electric field), or a moving charged object (magnetic field).
Force A p ush or p ull, either by an obvious agent or an invisible influence like gravity, electricity, or magnetism. In Newtonian
p hy sics, forces cause changes in motion.
Frame of reference The surroundings that share one’s state of motion; the p hy sical setting that establishes one’s p oint of
view for making measurements.
Free fall, free float Terms ap p lied to a situation in which an object moves under the influence of gravity alone.
Frequency The number of times p er second that a p eriodic occurrence, such as the oscillation of a wave, takes p lace.
Future In relativity, the future of an event consists of all those events that the given event can influence.

Galilean relativity The p rincip le, known since the time of Galileo and Newton, that the laws of motion p rovide no way to
distinguish different states of uniform motion. In other words, the question, Am I moving? is meaningless as far as the laws
of motion are concerned.
General relativity Einstein’s 1916 theory that describes gravity as the curvature of sp acetime.
Geodesic The straightest p ossible p ath through a sp ace or sp acetime that may itself have curvature. Great circles are the
geodesics on Earth’s surface.
Gravitational lens Any massive body whose gravitation, or sp acetime curvature in general relativity, bends light from more
distant objects. The result may be brightened, distorted, or multip le images.
Gravitational time dilation The p henomenon whereby time p asses more slowly near a gravitating object. Also called
gravitational redshift.
Gravitational wave A wave disturbance in the structure of sp acetime itself. Gravitational waves originate in the acceleration
of massive objects and p rop agate at the sp eed of light.

Hawking radiation A p henomenon associated with quantum p hy sics in the p resence of a black hole, whereby p article–
antip article p airs come sp ontaneously into existence just outside the hole. When one of the p air falls into the hole, the other
is left alone and ap p ears as if it were radiation emerging from the hole. Eventually, Hawking radiation sap s black holes of
their mass-energy.

Interference A wave p henomenon whereby two waves meeting at the same p lace simp ly add. In constructive interference, the
wave disturbances are in the same direction (e.g., crests meet crests, troughs meet troughs) and the effect is a strengthened
wave. In destructive interference, crests meet troughs and the overall wave is diminished.
Inertial reference frame Any reference frame in which the law of inertia is obey ed. In p ractice, the only inertial frames are
those in free fall (or free float).

Length contraction The reduction in the length of an object as measured by an observer with resp ect to whom the object is
moving.
Light A form of electromagnetic radiation visible to the human ey e. M ore loosely, in this book, the term is often used to
designate any electromagnetic radiation.
Light-year The distance light travels in 1 y ear.
Luxon A p article that travels at exactly the sp eed of light relative to any uniformly moving reference frame. The p hoton, or
quantized bundle of electromagnetic wave energy, is one familiar examp le.

Maxwell’s equations The four equations develop ed in the 1860s by James Clerk M axwell that describe all p henomena of
classical electricity and magnetism.
Mechanics The study of motion, one of the major branches of p hy sics.
Medium The substance of which a given wave is a disturbance. Air, for examp le, is the medium for sound waves, and water
for water waves. Light waves do not require a medium.
Michelson interferometer A device that uses interference of light waves traveling on two p erp endicular p aths to detect
minute changes in sp eed or distance between the two p aths.
Michelson–Morley experiment The famous 1887 exp eriment in which M ichelson and M orley failed in their attemp t to
detect Earth’s motion through the ether, desp ite using a M ichelson interferometer more than adequate to the task. The
failure of this exp eriment p aved the way for Einstein’s relativity.
Multiverse A multibranched sy stem of multip le universes, p rop osed by some cosmologists as rep resenting the overall
structure of all that exists.
Muons Subatomic p articles, created by cosmic ray s high in the atmosp here and heading Earthward at nearly the sp eed of light,
whose radioactive decay confirms time dilation.

Neutron star An object resulting from the collap se of a massive star and comp osed almost entirely of neutrons, with mass
about that of the Sun’s crammed into a sp here only a few miles across.
Newton’s laws of motion Three laws describing the relation between motion and force. The first states that an object not
subject to any force continues in uniform motion. The second states that an object’s acceleration is p rop ortional to the
force ap p lied to it and that for a given force the acceleration is less for a greater mass. The third law states that forces come
in p airs; if one object exerts a force on another, then the second exerts a force of equal strength, but op p osite direction, back
on the first object.

Orbit The p ath described by an object moving under the influence of gravity alone.

Past In relativity, the p ast of an event consists of all those events that can influence the given event.
Photon A p articlelike bundle of energy that, in quantum p hy sics, is associated with electromagnetic waves.
Planck length The tiny length—about 10-35 meter—at which quantum p hy sics should affect the nature of sp ace itself.
Planck time The tiny time—about 10-43 second—at which quantum p hy sics should affect the nature of time itself. To study
hap p enings at the Planck time or length scale will require a successful merging of quantum p hy sics with general relativity.
Positron Antip article to the electron.
Precession The gradual shift in the orientation of a p lanet’s orbit. General relativity p redicts that p recession should occur but
Newtonian gravitational theory does not. Precession has been detected in M ercury ’s orbit and in binary p ulsars.
Present In relativity there is no such thing as a universal p resent, which would consist of events near and far that are
simultaneous with what’s hap p ening here and now. Instead, the p resent of a given event consists, strictly sp eaking, of that
event alone.
Principle of Equivalence The statement that it is imp ossible to distinguish the effect of gravity from acceleration and the
absence of gravity from free fall.
Principle of Relativity The statement that the laws of p hy sics do not dep end on one’s frame of reference. Thus absolute
motion is a meaningless concep t; only relative motion is meaningful.
Proper length The length of an object as measured in a frame of reference in which the object is at rest.
Pulsar A rap idly sp inning neutron star that emits a searchlightlike beam of electromagnetic radiation, observed as a sequence
of p ulses as the beam sweep s rep eatedly by Earth.

Quantum gravity A theory —as y et undevelop ed—that would successfully combine quantum p hy sics with general relativity.
Quasar A very distant astrop hy sical object emitting colossal amounts of energy. Quasars are believed to be the central regions
of distant and therefore very early galaxies that emit cop ious radiation as material swirls into a massive black hole.

Retrograde motion The occasional reversal of direction that p lanets undergo when their motion through the sky is viewed
from Earth.

S imultaneous events Events that occur at the same time. In relativity, events that are simultaneous in one frame of reference
need not be simultaneous in other frames that are moving relative to the first.
S pacetime The union of sp ace and time into a single four-dimensional structure.
S pacetime interval The four-dimensional “distance” between events in sp acetime. The interval has the same value regardless
of one’s frame of reference.
S pecial relativity Einstein’s 1905 theory based on the p rincip le that the laws of p hy sics are the same in all reference frames
in uniform motion. The restriction to uniform motion is what makes this the special theory.
S tandard model The theory that describes the different elementary p articles and their interactions.
S tring theory A contemp orary theory that envisions fundamental entities not as p articles but as loop like strings. What we
consider elementary p articles would be different modes of vibration of these strings. String theory may have the p otential
 for unify ing quantum p hy sics with general relativity, p roviding a theory of every thing.

Tachyon A hy p othetical p article cap able of traveling faster than the sp eed of light, but not at light sp eed or slower. Tachy ons
have not been detected and their existence might wreak havoc on traditional notions of cause and effect.
Tardyon Any p article that travels at less than the sp eed of light relative to any uniformly moving reference frame. All known
matter is comp osed of tardy ons.
Time dilation A relativistic effect wherein the time between two events is shorter on a clock p resent at both events than it is
when measured by two sep arate clocks in a reference frame where the events take p lace at different p ositions. Time dilation
is sometimes described with the p hrase “moving clocks run slow,” but this is relativistically incorrect language for reasons
described in the text.
Twins paradox A p henomenon resulting because time dilation occurs on both the outbound and return legs of a round-trip
journey. A twin who makes such a journey returns to her starting p oint y ounger than her stay -at-home twin.

Uniform motion M otion in a straight line at constant sp eed.

Universal gravitation The idea, first develop ed by Newton, that every object in the Universe attracts every other object with
a gravitational force that dep ends on the objects’ masses and on the distance between them.

Wave A traveling disturbance that transp orts energy but not matter.
Wavelength The distance between adjacent wave crests.
Worldline The continuous p ath of an object in sp acetime, tracing out the entire history of the object.
Wormhole A hy p othetical “bridge” that would constitute a shortcut between two otherwise distant regions of sp acetime, or
even to another universe.
FURTHER READINGS
• • •

Brian, Denis. Einstein: A Life. New York: John Wiley & Sons, 1996.
A modern and thoroughly candid account of Einstein’s life, among the first to exp loit the Einstein Archives and Einstein
documents made widely available in the 1980s.
Gott, J. Richard. Time Travel in Einstein’s Universe: The Physical Possibilities of Travel through Time. Boston: Houghton
M ifflin, 2001.
Gott, a Princeton p hy sicist, not only elaborates relativity ’s clear imp lication that time travel to the future is p ossible but
also exp lores the more sp eculative p rosp ect of time travel to the p ast—a p rosp ect that some p hy sicists find increasingly
worthy of serious study.
Greene, Brian. The Elegant Universe: Superstrings, Hidden Dimensions, and the Quest for the Ultimate Theory. New York: W.
W. Norton, 1999.
Columbia University theorist Brian Greene gives a thorough and thoroughly readable account of string theory, the leading
candidate for a “theory of every thing” that would unite relativity and quantum p hy sics. Greene’s own contributions to the
field are substantial, and he writes with an insider’s deep knowledge and y et with the skill of one who can communicate
comp lex ideas to nonsp ecialists.
Guth, Alan H. The Inflationary Universe: The Quest for a New Theory of Cosmic Origins. Reading, M A: Addison Wesley,
1997.
Guth is one of the originators of the inflationary universe idea, which resolves many of the conundrums raised by the Big
Bang theory. This lively, accessible, y et authoritative account of our modern understanding of the origin of the Universe
mixes solid science with the history of and p ersonalities at the forefront of cosmology.
Hafele, J. C., and R. E. Keating. “Around the World Atomic Clocks: Predicted Relativistic Time Gains” and “Around the
World Atomic Clocks: Observed Relativistic Time Gains.” Science 177 (July 1972): 166–70.
This account of a real “twins” exp eriment done with highly accurate atomic clocks leaves no doubt about the reality of time
dilation.
Livingston, Dorothy M ichelson. The Master of Light. New York: Charles Scribner’s Sons, 1973.
This biograp hy by M ichelson’s daughter p rovides an insightful account of the M ichelson–M orley exp eriment and its
history, but also sp ares no detail of M ichelson’s often turbulent p ersonal and p rofessional life.
Pais, Abraham. Einstein Lived Here. New York: Oxford University Press, 1994.
A thematic, nonmathematical introduction to Einstein’s life and thought. Author Abraham Pais, a p hy sicist and Einstein
scholar, has written extensively about Einstein’s science and life. His earlier book, Subtle Is the Lord: The Science and the
Life of Albert Einstein (Oxford University Press, 1982), delves deep ly and quantitatively into Einstein’s scientific work.
Paterniti, M ichael. Driving Mr. Albert. New York: Dell Publishing, 2000.
Ever wonder what hap p ened to Einstein after his death? His brain, at least, goes on—sloshing about in a Tup p erware
container! This amusingly imp robable travelogue chronicles the true story of a cross-country drive with Einstein’s brain
and its eccentric steward.
Py enson, Lewis. The Young Einstein. Boston: Adam Hilger Ltd., 1985.
A scholarly account of Einstein from childhood through the develop ment of both relativity theories; p articularly strong on
the cultural and scientific milieu in which Einstein arose and on scientists’ and mathematicians’ reactions to Einstein’s
work.
Smolin, Lee. Three Roads to Quantum Gravity. New York: Basic Books, 2001.
A self-confessed op timist, p hy sicist Lee Smolin believes we are only y ears away from unify ing general relativity and
quantum p hy sics to p roduce a “theory of every thing.” In this lively and very contemp orary book, he outlines three
ap p roaches that show p romise.
Tay lor, Edwin F., and John Archibald Wheeler. Spacetime Physics, 2d ed. New York: W. H. Freeman, 1992.
Written for sop homore-level college p hy sics students, this book p resents sp ecial relativity in a lively and entertaining sty le
that emp hasizes the underly ing simp licity and fundamental p rincip les of the theory. Some sections of the book are
reasonably math-free and p rovide deep insights into relativity. If y ou don’t mind a little algebra, Spacetime Physics will
greatly enhance y our understanding of Einstein’s theory.
Thorne, Kip . Black Holes and Time Warps: Einstein’s Outrageous Legacy. New York: W. W. Norton, 1994.
This definitive history of black holes and related p henomena of general relativity will convince y ou that black holes really
do exist. Written in the first p erson by a leading relativity researcher, the book will reward a p ersevering reader with a really
thorough but nonquantitative understanding of black holes in p articular and relativity in general.
Will, Clifford. Was Einstein Right? Putting General Relativity to the Test. Basic Books, A Division of Harp erCollins, 1986.
This highly readable book gives a thorough look at both the classical and fairly contemp orary tests of general relativity.
Written for the nonscientist, it’s a fascinating blend of p hy sics, p ersonalities, and history.
Wambsganss, Joachim. “Gravity ’s Kaleidoscop e.” Scientific American 285 (November 2001): 64–71.
A nice account of gravitational lenses that shows how they work and the multip le uses to which astrop hy sicists are now
p utting them. Great illustrations, too!
INDEX
• • •

aberration of starlight, 66
acceleration, 30
accretion disk, 223
action at a distance, 45, 175, 193
Amp ère, André-M arie, 43
antimatter, 156
antip article, 156
Apollo 13, 184
Aristotle, 20

beam sp litter, 72
Bell, John, 161
Big Bang, 3, 231
binary neutron star, 207
binary p ulsar, 207
gravitational waves from, 227
binary star, 57
containing black hole, 223
Biot, Jean-Bap tiste, 43
black hole, 10, 206, 218–25
in binary star sy stem, 223
in galactic centers, 224
p rop erties, 218
Brahe, Ty cho, 23

causality, 142–43
clock, 100
atomic, 7, 121, 215
constructive interference, 39
Cop ernicus, Nicolaus, 13, 22, 63
cosmic microwave background, 233
cosmological constant, 230
cosmology, 229
Coulomb, Charles Augustin, 42
critical density, 232

dark energy, 233

dark matter, 232
destructive interference, 39
Dimoup olos, Savas, 237
Dop p ler effect, 208
double star, 57
E=mc2, 154
Earth
motion relative to ether, 62
Eddington, Arthur, 211
Einstein, Albert
association with nuclear weap ons, 158
childhood y ears, 78
college y ears, 79
and EPR p aradox, 161
“greatest blunder,” 231
1905 p ap ers, 81
Nobel Prize, 81
electric charge, 42, 171
electric current, 42
electric field, 47
electricity
static, 42
electromagnetic induction, 44
electromagnetic sp ectrum, 51
electromagnetic wave, 49
sp eed of, 50
electromagnetism, 43
consistency with sp ecial relativity, 175
electron, 44
elsewhere, 143
energy
conversion to matter, 156
equivalent to matter, 155
kinetic, 159
and momentum, 171
in wave, 37
ep icy cle, 21
EPR p aradox, 161
escap e sp eed, 21
ether, 56
p rop erties of, 56
ether drag, 64
Euclidean geometry, 190
event, 96
on sp acetime diagram, 145
event horizon, 218
exp eriment, definition of, 14

Faraday, M ichael, 44
field
electric, 47
gravitational, 45
magnetic, 47
Fitzgerald, George, 77
force, 20
frame of reference, 32
inertial, 177
Franklin, Benjamin, 42
free fall, 184
free float, 184
frequency
of electromagnetic wave, 51
Friedmann, Aleksandr, 230
future, 140

Galilean relativity, 32
Galileo, 14, 24
and Tower of Pisa, 177
general theory of relativity, 174
geodesic, 195
geometry
Euclidean, 190
in general relativity, 192
Global Positioning Sy stem, 216
Gott, J. Richard, 4
GPS. See global p ositioning sy stem
gravitational field, 45
gravitational lens, 9, 212–14
gravitational redshift, 215
gravitational time dilation, 215
gravitational waves, 225
detection of, 226
graviton, 160
gravity, 27
Einstein’s view, 192
in general relativity, 192
Newton’s view, 28
strong, 218
weak, 203
great circle, 191
Guth, Alan, 6

Hawking, Step hen, 220, 236

Hawking radiation, 236
Hertz, Heinrich, 49
Hubble, Edwin, 231
Hubble Sp ace Telescop e, 9
Hulse, Russell, 207
Huy gens, Christiaan, 39

inertia, 159
law of, 177
inertial reference frame, 177, 186
inflation, 237
interference, 39
interferometer, M ichelson, 70

Jup iter, 24

Kep ler, Johannes, 23

kinetic energy, 159

law of inertia, 177

length contraction, 115
Li, Li-Xin, 4
light
bending of, 209–12
as electromagnetic wave, 51
sp eed of. See sp eed of light
wave theory, 39
lightlike interval, 169
light-y ear, 109
LIGO, 226
Linde, Andrei, 5
LISA, 226
Lorentz, H. A., 77

magnetic field, 47
magnetism, 42
relation to electricity, 172
M anhattan p roject, 159
M arconi, Guglielmo, 49
M ari´c, M ileva, 79
M ars Rover, 141
mass-energy
conservation of, 156
mass-energy equivalence, 155, 199
matter
conversion to energy, 156
equivalent to energy, 155
M axwell, James Clerk, 48
M axwell’s equations, 48
and ether, 60
mechanics, 31
medium, for wave, 36
M ercury, 204
M ichelson, Albert, 70
M ichelson interferometer, 70
M ichelson-M orley exp eriment, 70–76, 117
ap p aratus, 72
concep tual descrip tion, 70
Einstein’s knowledge of, 70
result, 76
microlensing, 214
microwave oven, 11
M inkowski, Hermann, 168
missing mass, 214
momenergy, 171
M orley, Edward, 75
motion
natural state, 20, 23, 26, 196
relative, 32, 91
retrograde, 21
study of, 19
uniform, 176
M ount Washington, 119
M -theory, 235
M ultiverse, 5, 237
muon, 119

neutrino, 160
neutron star, 206
Newton, Isaac, 14, 27
Newtonian gravitation
p roblems with, 175
Nobel Prize, 81
nuclear p ower
 and E=mc2, 157

observing
versus seeing, 105
Oersted, Hans Christian, 43
orbit
circular, 29
ellip tical, 23
of M ercury, 204-5

p air-creation, 156
Parable of the Survey ors, 165
p arallel universe, 237
p ast, 140
p erihelion, 205
p hoton, 160, 234
Planck length, 235
Planck time, 235
Podolsky, Boris, 161
p ositron, 156
p recession, 204
p resent, 139
Princip le of Equivalence, 178–83
p rincip le of Galilean relativity, 32
Princip le of Relativity, 81
and c as maximum sp eed, 149
imp lications for sp eed of light, 83
p rop er length, 116
Ptolemy, 21
p ulsar, 207

quantum gravity, 235

quantum p hy sics, 234
quasar, 212, 224
quintessence, 233

redshift, gravitational, 215

reflection, 39
refraction, 39
relativistic velocity addition, 152
retrograde motion, 21
Rosen, Nathan, 161

Sagan, Carl, 6, 238

satellite
artificial, 29
Savart, Félix, 43
seeing
versus observing, 105
simultaneity, 129–30
relativity of, 132
simultaneous events. See simultaneity
sound, 37
[false] analogy with light, 89
sp eed of, 37, 55, 59, 89
sp ace
commonsense notions, 95
relativity of, 94
sp acelike interval, 169
sp acetime, 143, 167
curvature, 192, 230, 232
geometry of, 190, 230, 232
sp acetime diagram, 145
sp acetime interval, 167
lightlike, 169
sp acelike, 169
timelike, 168
sp ecial theory of relativity, 82
sp eed of light, 50, 55
as conversion factor between sp ace and time, 170
exact value, 87
invariance of, 83, 88
in matter, 52
as maximum p ossible sp eed, 149–53, 163
numerical values, 52
relative to ether, 56
relative to source, 57
and relativity p rincip le, 83
in vacuum, 52
in water, 163
standard model, 237
static electricity, 42
string theory, 235
sunsp ots, 25
sup erluminal communication, 162
sup ernova, 140, 226, 233
survey ors p arable, 165
Szilard, Leo, 159

tachy on, 160

tardy on, 161
Tay lor, Edwin, 165
Tay lor, Josep h, 207
tennis, 11, 176
Theory of Every thing, 236
theory of invariance, 164
theory of relativity
alternate name for, 164
tidal forces, 190
time
commonsense notions, 95
Newton’s views on, 94
relativity of, 94
time dilation, 96–108
ap p lied to sp ace travel, 111
formula, 106
gravitational, 215
mathematical derivation, 241
muon exp eriment, 119
timelike interval, 168
time loop , 4
time travel, 8, 125–26, 220
Tower of Pisa, 177
twins p aradox, 122–25
in general relativity, 197
uniform motion, 26, 176
Universe
accelerated exp ansion, 233
exp ansion of, 231
fate of, 232
geometry of, 232
origin of, 3

Venus, 12

wave, 36
electromagnetic, 49
wavelength, 41
of electromagnetic wave, 51
weak gravity, 203
weightlessness, 184
Wheeler, John Archibald, 165
white dwarf, 217
worldline, 148
of Earth, 202
wormhole, 7, 238

Young, Thomas, 39
COPYRIGHT
• • •

Copyright © 2003 by Richard Wolfson

All rights reserved

Printed in the United States of America
First Edition

For information about permission to reproduce selections from this book,

write to Permissions, W. W. Norton & Company, Inc., 500 Fifth Avenue,
New York, NY 10110

Book design by Margaret M. Wagner

Production manager: Julia Druskin

Library of Congress Cataloging-in-Publication Data

Wolfson, Richard.
Simply Einstein : relativity demystified / Richard Wolfson.— 1st ed.
p. cm.
Includes bibliographical references and index.
ISBN 0-393-05154-4 (hardcover)
ISBN 978-0-393-24218-8 (e-book)
1. Relativity (Physics)—Popular works. I Title: Relativity demystified.
II. Title.
QC173.57 .W65 2003
530.11—dc21
2002002984

W. W. Norton & Company, Inc.

500 Fifth Avenue, New York, N.Y. 10110
www.wwnorton.com

W. W. Norton & Company Ltd.

Castle House, 75/76 Wells Street, London W1T 3QT