528 Chapter 12 Electrodynamics and Relativity

Problem 12.19 The parallel between rotations and Lorentz transformations is even
more striking if we introduce the rapidity:
θ ≡ tanh−1 (v/c). (12.34)

(a) Express the Lorentz transformation matrix (Eq. 12.24) in terms of θ, and
compare it to the rotation matrix (Eq. 1.29).
In some respects, rapidity is a more natural way to describe motion than ve-
locity.11 For one thing, it ranges from −∞ to +∞, instead of −c to +c. More
significantly, rapidities add, whereas velocities do not.

(b) Express the Einstein velocity addition law in terms of rapidity.

(ii) The invariant interval. The scalar product of a 4-vector with itself,
a μ aμ = −(a 0 )2 + (a 1 )2 + (a 2 )2 + (a 3 )2 , can be positive (if the “spatial” terms
dominate) or negative (if the “temporal” term dominates) or zero:
If a μ aμ > 0, a μ is called spacelike.
If a μ aμ < 0, a μ is called timelike.
If a μ aμ = 0, a μ is called lightlike.
Suppose event A occurs at (x 0A , x 1A , x 2A , x 3A ), and event B at (x B0 , x B1 , x B2 , x B3 ).
The difference,
μ μ
x μ ≡ x A − x B , (12.35)
is the displacement 4-vector. The scalar product of x μ with itself is called the
invariant interval between two events:
I ≡ (x)μ (x)μ = −(x 0 )2 + (x 1 )2 + (x 2 )2 + (x 3 )2 = −c2 t 2 + d 2 ,
(12.36)
where t is the time difference between the two events and d is their spatial sep-
aration. When you transform to a moving system, the time between A and B is
altered (t¯ = t), and so is the spatial separation (d̄ = d), but the interval I remains
the same.
If the displacement between two events is timelike (I < 0), there exists an
inertial system (accessible by Lorentz transformation) in which they occur at the
same point. For if I hop on a train going from (A) to (B) at the speed v = d/t,
leaving event A when it occurs, I shall be just in time to pass B when it occurs;
in the train system, A and B take place at the same point. You cannot do this for
a spacelike interval, of course, because v would have to be greater than c, and no
observer can exceed the speed of light (γ would be imaginary and the Lorentz
transformations would be nonsense). On the other hand, if the displacement is
spacelike (I > 0), then there exists a system in which the two events occur at the
same time (see Prob. 12.21). And if the displacement is lightlike (I = 0), then the
two events could be connected by a light signal.
11 E. F. Taylor and J. A. Wheeler, Spacetime Physics, 1st ed. (San Francisco: W. H. Freeman, 1966).
12.1 The Special Theory of Relativity 529

Problem 12.20

(a) Event A happens at point (x A = 5, y A = 3, z A = 0) and at time t A given by

ct A = 15; event B occurs at (10, 8, 0) and ct B = 5, both in system S.

(i) What is the invariant interval between A and B?

(ii) Is there an inertial system in which they occur simultaneously? If so, find
its velocity (magnitude and direction) relative to S.

(iii) Is there an inertial system in which they occur at the same point? If so, find
its velocity relative to S.

(b) Repeat part (a) for A = (2, 0, 0), ct = 1; and B = (5, 0, 0), ct = 3.

Problem 12.21 The coordinates of event A are (x A , 0, 0), t A , and the coordinates of
event B are (x B , 0, 0), t B . Assuming the displacement between them is spacelike,
find the velocity of the system in which they are simultaneous.

(iii) Space-time diagrams. If you want to represent the motion of a particle

graphically, the normal practice is to plot the position versus time (that is, x runs
vertically and t horizontally). On such a graph, the velocity can be read off as the
slope of the curve. For some reason, the convention is reversed in relativity: every-
one plots position horizontally and time (or, better, x 0 = ct) vertically. Velocity
is then given by the reciprocal of the slope. A particle at rest is represented by a
vertical line; a photon, traveling at the speed of light, is described by a 45◦ line;
and a rocket going at some intermediate speed follows a line of slope c/v = 1/β
(Fig. 12.21). We call such plots Minkowski diagrams.
The trajectory of a particle on a Minkowski diagram is called a world line.
Suppose you set out from the origin at time t = 0. Because no material object
can travel faster than light, your world line can never have a slope less than 1.
Accordingly, your motion is restricted to the wedge-shaped region bounded by
the two 45◦ lines (Fig. 12.22). We call this your “future,” in the sense that it is the

Your future at t

Your future,
ct at t = 0
Rocket Photon

Present Present
x Your world line

Particle
at rest Your past,
at t = 0

FIGURE 12.21 FIGURE 12.22

530 Chapter 12 Electrodynamics and Relativity

locus of all points accessible to you. Of course, as time goes on, and you move
along your chosen world line, your options progressively narrow: your “future”
at any moment is the forward “wedge” constructed at whatever point you find
yourself. Meanwhile, the backward wedge represents your “past,” in the sense
that it is the locus of all points from which you might have come. As for the
rest (the region outside the forward and backward wedges), this is the generalized
“present.” You can’t get there, and you didn’t come from there. In fact, there’s no
way you can influence any event in the present (the message would have to travel
faster than light); it’s a vast expanse of spacetime that is absolutely inaccessible
to you.
I’ve been ignoring the y and z directions. If we include a y axis coming out of
the page, the “wedges” become cones—and, with an undrawable z axis, hyper-
cones. Because their boundaries are the trajectories of light rays, we call them the
forward light cone and the backward light cone. Your future, in other words,
lies within your forward light cone, your past within your backward light cone.
Notice that the slope of the line connecting two events on a space-time dia-
gram tells you at a glance whether the displacement between them is timelike
(slope greater than 1), spacelike (slope less than 1), or lightlike (slope 1). For
example, all points in the past and future are timelike with respect to your present
location, whereas points in the present are spacelike, and points on the light cone
are lightlike.
Hermann Minkowski, who was the first to recognize the full geometrical sig-
nificance of special relativity, began a famous lecture in 1908 with the words,
“Henceforth space by itself, and time by itself, are doomed to fade away into
mere shadows, and only a kind of union of the two will preserve an independent
reality.”12 It’s a lovely thought, but you must be careful not to read too much into
it. For it is not at all the case that time is “just another coordinate, on the same
footing with x, y, and z” (except that for obscure reasons we measure it on clocks
instead of rulers). No: Time is utterly different from the others, and the mark of its
distinction is the minus sign in the invariant interval. That minus sign imparts to
spacetime a hyperbolic geometry that is much richer than the circular geometry
of 3-space.
Under rotations about the z axis, a pointP in the x y plane describes a circle:
the locus of all points a fixed distance r = x 2 + y 2 from the origin (Fig. 12.23).
Under Lorentz transformations, however, it is the interval I = (x 2 − c2 t 2 ) that is
preserved, and the locus of all points with a given value of I is a hyperbola—or,
if we include the y axis, a hyperboloid of revolution. When the displacement is
timelike, it’s a “hyperboloid of two sheets” (Fig. 12.24a); when the displacement
is spacelike, it’s a “hyperboloid of one sheet” (Fig. 12.24b). When you perform a
Lorentz transformation (that is, when you go into a moving inertial system), the
coordinates (x, t) of a given event will change to (x̄, t¯), but these new coordinates
will lie on the same hyperbola as (x, t). By appropriate combinations of Lorentz
transformations and rotations, a spot can be moved around at will over the surface

12 A. Einstein et al., The Principle of Relativity (New York: Dover, 1923), Chapter V.
12.1 The Special Theory of Relativity 531

P
r
x

FIGURE 12.23

ct ct

(a) (b)

FIGURE 12.24

of a given hyperboloid, but no amount of transformation will carry it, say, from
the upper sheet of the timelike hyperboloid to the lower sheet, or to a spacelike
hyperboloid.
When we were discussing simultaneity, I showed that the time ordering of two
events can, at least in certain cases, be reversed, simply by going into a mov-
ing system. But we now see that this is not always possible: If the displacement
4-vector between two events is timelike, their ordering is absolute; if the interval
is spacelike, their ordering depends on the inertial system from which they are
observed. In terms of the space-time diagram, an event on the upper sheet of a
timelike hyperboloid definitely occurred after (0, 0), and one on the lower sheet
certainly occurred before; but an event on a spacelike hyperboloid occurred at
positive t, or negative t, depending on your reference frame. This is not an idle
curiosity, for it rescues the notion of causality, on which all physics is based. If
it were always possible to reverse the order of two events, then we could never
say “A caused B,” since a rival observer would retort that B preceded A. This
embarrassment is avoided, provided the two events are timelike or lightlike sepa-
rated. And causally related events are—otherwise no influence could travel from
one to the other. Conclusion: The displacement between causally related events is
always timelike, and their temporal ordering is the same for all inertial observers.