An Intuitive Approach to Cosmic Horizons

Adam Neat

Citation: The Physics Teacher 57, 80 (2019); doi: 10.1119/1.5088465

View online: https://doi.org/10.1119/1.5088465
View Table of Contents: https://aapt.scitation.org/toc/pte/57/2
Published by the American Association of Physics Teachers
An Intuitive Approach to Cosmic
Horizons
Adam Neat, Swarthmore College, Swarthmore, PA

H
ow far, in space, can we see? And can we see an object be redshifted out of existence, i.e., an infinite amount?
whose Hubble recessional velocity exceeds the speed A related question about our observational limits is, “Giv-
of light? Maybe you’ve thought about these questions en that the universe is finite in age, how far, in space, can we
before, or perhaps you’ve seen them discussed in the literature see?” One might reason: If the universe is 14 billion years old,
or mentioned in the media. With the recent popularity of and light travels at a constant rate of c, the greatest distance a
inflation and Big Bang cosmology, they’re hard to avoid. The photon can have traveled since the beginning of time is 14 bil-
discussion that follows is an attempt to resolve some common lion light-years. This must therefore be the greatest distance
misconceptions—often seen in the popular literature—con- we can see (assuming perfect transparency). There may exist
cerning the above two questions, and to do so in a way that ap- more distant objects, but their light will not have reached us
peals to kinematical intuition. A simple thought experiment yet.
will be used to initiate the discussion and to answer the ques- To restate the above two conclusions: we cannot see objects
tion, “Can we see objects with faster-than-light recessional receding from us faster than light, and we cannot see anything
velocity?” Hubble’s law, along with a simple assumption about farther away than c3age of universe. It’s not uncommon to
the kinematics of light in expanding space, will be used to de- find conclusions, or reasoning, similar to that above in intro-
rive expressions, customarily derived in a general relativistic ductory-level accounts or physics-for-the-layperson descrip-
context, that allow cosmologists to determine our observa- tions.3-14 Both conclusions, however, turn out to be wrong.
tional limits and define our cosmological horizons. Some of
the results may surprise you. Before we delve into the topic Thought experiment: Seeing a faster-than-
fully, though, let’s first lay some theoretical groundwork. light object
Distant objects travel away from us faster than the speed To help us get a better handle on why it is that we can, or
of light, and it’s no violation of special relativity. This strange can’t, see faster-than-light objects, consider the following
fact is a consequence of one of modern cosmology’s most thought experiment. (This is what got me thinking about
elementary postulates—space is expanding. The distance be- cosmic horizons in the first place.) Three observers, A, B, and
tween two vastly separated objects in our universe increases C, are separated linearly in space, such that the Hubble reces-
with time, not because the objects are moving through space, sional velocity between A and C is greater than light speed,
but because the space in between them is expanding. In a uni- but the recessional velocities between A and B, and B and C
verse where space expands uniformly, the Hubble recessional are less than light speed (see Fig. 1).
velocity (recessional velocity due merely to spatial expansion)
is related to the distance between objects by the following very
simple equation, often referred to as Hubble’s law1: A B C
vr = HD, (1)   
Vr A,B < light speed Vr B,C < light speed
where vr is recessional velocity, D is separation distance,2 and
H is the Hubble constant. Two major implications of this law Vr A,C > light speed
are (1) recessional speed increases linearly with distance at
Fig.1. Three observers separated in space. The distances between
any one moment in time; and (2) there is no limit to an ob- observers is such that the recessional velocity (vr) is greater than
ject’s recessional speed. Not only is faster-than-light recession light speed for A and C, but less than light speed for A and B, and
not prohibited, it’s what all matter does outside a radius of B and C.
D = c/H.
A natural question that may arise in response to the above Also—and this is important—let’s assume, for simplicity,
two points is, “If something is traveling away from me faster that the recessional velocities are constant in time. Accord-
than light (FTL), can I see it?” If the object never slows down, ing to the logic in the preceding section, observers A and C
one may reason, how can light from me ever catch up with should not be able to see each other, as vr exceeds c. Howev-
that object; and, by symmetry, how can light from that object er, there’s no reason why A shouldn’t be able to see B, and B
ever catch up with me? Reasoning this way, FTL objects seem shouldn’t be able to see C. Now, suppose A sends a photon to
to be beyond our observational reach. To reinforce this con- B, and instead of intercepting the photon B lets it pass by. Will
clusion, consider the effects of redshift. Our special relativistic A’s photon eventually reach C? If we answer “no,” we must ex-
intuition might tell us that redshift approaches infinity as the plain why this “passing by” photon should behave any differ-
recessional velocity of the emitter approaches the speed of ently than a photon that originates directly from B, because as
light. If light from an FTL object could reach us, would it not we previously established B and C can exchange photons. But

80 THE PHYSICS TEACHER ◆ Vol. 57, February 2019 DOI: 10.1119/1.5088465

if we answer “yes,” we’ve contradicted our original conclusion result, no object receding from us at constant or decreasing
above—that light can’t be exchanged between two FTL ob- speed is beyond our observational reach. Light from us will
jects. So what’s the solution? always catch up with such objects and, by symmetry, light
The answer to the above question is yes. The photon from from such objects will always catch up with us. This, I think,
A will reach C. If light from B is always able to reach C, and is the most intuitive way to explain why it is that we can see
light from A is always able to reach B, light from A will always faster-than-light receding objects. However—and this may
reach C. And no, this light will not be redshifted an infinite have occurred to the reader—if the regions of space also ac-
amount. A photon from A will redshift a finite amount as it celerate apart, due to accelerating spatial expansion, the light
travels to B and another finite amount as it travels from B to might not catch up with the object. We’ll discuss this situation
C. (The factor by which the photon’s wavelength increases shortly.
during some time interval is equal to the factor by which the
universe expands during that time interval.) How far and “how fast” can we see?
So what’s wrong with our original argument above? How Understanding, explicitly, how light recedes from its
can a photon from observer A catch up with C if C recedes source will provide insight into our observational limits.
from A faster than light? The flaw in this question and in The time dependent distance of a photon, with respect to its
our first argument above, surprisingly, turns out to be the source, in non-expanding space is, of course, Dlight = ct, where
assumption that light recedes from its source at the constant t is travel time. In expanding space, because the recessional
rate of c. In a universe where space expands, light does not speed of light is changing—increasing with distance accord-
recede from its source at a constant speed, but in fact must ac- ing to Eq. (2)21—to determine Dlight we must set up and solve
celerate. When the light from A reaches B, it will be receding the appropriate differential equation. Equation (2) allows us
from A not at the speed of light but at the speed of light plus to do that. Written in differential form, Eq. (2) becomes
the recessional velocity of B.
To be more explicit, consider again Eq. (1), Hubble’s law.
The recessional speed of some region of space with respect (3)
to us is HD. A photon traveling through space, radially away for photon motion through space away from the emitter.
from us, within this region will therefore have a recessional Solving this gives the photon distance from its source as a
speed of c plus HD.15-18 A good analogy is that of an ant function of time; we just need to know the time dependence
walking on top of an expanding rubber band. The ant’s veloci- of H.
ty with respect to a point p on the rubber band, some distance The Hubble “constant” H is, of course, not a constant in
away from the ant, is the ant’s walking velocity with respect to time but a constant in space—it’s presumed to be the same
the rubber, plus the recessional velocity of the region of rub- everywhere on cosmological scales. In a space where objects
ber upon which the ant is standing, with respect to the point recede from one another at constant velocities, H is equal
p. In a similar fashion, for a universe of flat spatial geome- to 1/T, where T is the age of the universe. More generally
try,17 photon recessional speed with respect to an object some though, H is recessional velocity per unit distance. And even
distance D away is the sum19 of the recessional speed due to slightly more generally, H is the rate at which the scale of space
spatial expansion HD and the speed of light c. Or, changes divided by the scale itself. In cosmology, the state of
the universe’s expansion (accelerating, decelerating, constant,
vlight = HD  c , (2)
shrinking, etc.) is represented by the time-dependent scale
where a positive c indicates photon motion through space factor of space, R(t). When the scale of space increases lin-
radially away from the observer, and a negative c, radially early with time, R(t) = bt, and recessional velocities remain
towards. This equation, like Eq. (1), holds even when the constant. When R(t) increases with positive inflection, the
expansion rate of space is not constant.20 Though the above expansion of space accelerates, and when R(t) increases with
result is based only on our intuition that an object’s net re- negative inflection, the expansion decelerates (see Fig. 2).
cessional rate is the sum of its velocity through space and its And, of course, R(t) = b means . no expansion at all. The Hub-
recessional velocity due to spatial expansion, Eq. (2) is per- ble term H is defined to be R/R, or the rate at which the scale
fectly consistent with general relativity and can be derived of space changes divided by the scale itself.22 In the linearly
from cosmological principles within a GR context.15-18 [Note: expanding, constant-recessional-velocity
. universe, where
It’s important for the reader to keep in mind that Eq. (2), and R(t) = bt, H = R /R = 1/T. Figure 2 shows how linear expansion
the assumptions leading up to it, are based on the expanding compares with two other (more realistic) models—one in
space interpretation of cosmological redshift, derived from which spatial expansion continually slows with time [R(t)~
general relativity. For a nice discussion of some of the empiri- t2/3], and one in which the scale factor undergoes an inflec-
cal evidences supporting spatial expansion, see Ref. 18.] tion point, causing the rate of spatial expansion to go from
So, because light always travels through space at the con- decelerating to accelerating [R(t)~ sinh2/3 t].23
stant rate of c, and each region of space recedes from us at a The “scale of space” is a rather abstract notion. To relate
rate proportional to distance, the recessional velocity of light, this to something more concrete, imagine two distant galax-
with respect to the emitter, must continually increase. As a ies at rest with respect to the same frame. If the scale of space

THE PHYSICS TEACHER ◆ Vol. 57, February 2019 81

 2
A more general approach for using Eq. (3) to solve for the
. H with its explicit time-de-
photon distance is not to replace
R(t) sinh 3 t
pendent expression, but with R /R, its definition. The equa-
R(t) t
tion of motion for a photon receding from its source then
R(t) t 2 /3
becomes


(6)

And solving this yields22

(7)
0.0

0.0
where t0 and te are the reception and emission times, respec-
Time tively, for an exchanged photon, and Dlight is the distance of
Fig. 2. Scale of space as a function of time for three different the photon at time t0 , or D0. The conventional way to derive
expansion models. Non-accelerating (linear) expansion is repre- Eq. (7)— in fact, the only way I’ve seen in the literature— uses
sented by R(t) ~ t. concepts typically found within a general relativistic context,
namely, the FLRW metric (Friedmann–Lemaître–Robert-
were to increase by a factor of n, the distance between the two son–Walker), and the idea that light travels along null geode-
galaxies would change by a factor of n, even though the galax- sics.15,16,19,25 Terms such as comoving coordinates, spacetime
ies remain at rest with respect to the same frame. The curves metric, and proper time and distance require explanation,
in Fig. 2 represent how distances between galaxies change as well as how these ideas relate to spacetime geometry and
with time solely due to spatial expansion, for three different photon trajectories. For students not acquainted with general
expansion rates. relativity, the method outlined above should be, conceptually,
So, for the constant-recessional-velocity universe, we re- much more transparent.26
place H in Eq. (3) with 1/T, yielding
The particle horizon
. (4) The distance of a receding photon at time t0 [Eq. (7)] is al-
so the distance of an object that intercepts the photon at time
Replacing T with te + t, where te is the age of the universe at t0. Therefore, Eq. (7) gives the distance between two objects
the time of emission and t is the travel time of the photon, and at the time of light reception, D0, in terms of the emission and
solving yields the following expression for the distance of a reception time of an exchanged photon.
photon from the source of emission: To compare and contrast: In non-expanding space the
distance between two objects that reside in the same frame
(5) of reference, in terms of light emission and reception time, is
simply D0 = c(t0 – te); but in a linearly expanding model, D0
Notice how this equation differs from the non-expand- = ct0 ln(t0/te). In non-expanding space D0 maxes out at ct0 ,
ing-space form, with which we are all familiar, Dlight = ct. In consistent with the statement that the most distant observable
an expanding universe, the distance between a photon and object can be no farther than c 3age of universe away. But in a
the emitter depends not only on how long the photon has spatially expanding universe of constant recessional velocity,
been traveling, t, but on the age of the universe at the time as you can check, D0 has no limit27 (quite a bit farther than
of emission, te. This is because the recessional velocity per c 3age of universe!). Meaning, in such a universe, our causal
unit distance is higher the farther back in time we go, which radius would extend to infinity, and we would be able to “see”
is not due to the expansion rate being higher (remember, the things28 that are currently located at any distance—assuming
rate is constant in this example, i.e., neither accelerating nor perfect transparency.
decelerating) but rather due to the mutually receding regions When emission time te is set to 0 in Eq. (7), and t0 = T,
of space being more. tightly packed together. So, H is higher at where T is the age of the universe, we have an expression
earlier times (H = R/R, and R is smaller at earlier times while cosmologists call the particle horizon, DPH .15,16,19,25 It rep-
.
R is constant), but the scale of space changes at the same rate. resents the current location of the most distant object we can
Substituting see. The size of the “observable” universe is often equated
with this distance.19 One can also think of this distance as our
causal radius29 as alluded to above. For the R(t) = bt universe,
interestingly, DPH = `. However, for universes where the ex-
with its series expansion, one can show that Eq. (5) reduces pansion rate continually slows down with time, DPH is finite.
to Dlight = ct for travel times much smaller than the age of the For instance, if R(t) = bt2/3 [a “matter-dominated” universe,
universe, t << te. where the scale of space increases more slowly as time pro-

82 THE PHYSICS TEACHER ◆ Vol. 57, February 2019

 gresses (see Fig. 2)], as you can check, the particle horizon is greater emission distance, in spatially expanding universes.
3c3age of universe. The important point to remember here is The greatest emission distance, De,max (what some may call
that one’s causal radius (the particle horizon) depends on how the greatest distance we can see), in non-expanding space is,
space expands, R(t). What’s the most distant object we can of course, c 3age of universe. In expanding space the greatest
see? It may be quite a bit farther than c3age of universe away; emission distance, like the particle horizon, is a model-de-
exactly how far depends on how space expands. According pendent quantity. In the linearly expanding model, De,max,
to the most widely accepted model (ΛCDM), our particle interestingly, turns out to be cT/e , where e is the base of the
horizon is about 3.4c3age of universe, or about 47 billion natural logarithm, 2.71828... . See Fig. 4.
light-years, away.18 Figure 3 shows how the particle horizon The slope of the curves in Fig. 4 represents the velocity of a
depends on expansion rate for various R(t) ~ t n universes (re- particle of light—received today—with respect to us, over the
ferred to as “power law” models20). When n is between 0 and course of its history [recall from Eq. (2) that this is HD –c].
1, expansion slows with time. For such universes, Eq. (7) gives Light from the Big Bang that we see today initially receded
DPH = cT/ (1 – n). from us very quickly due to spatial expansion (as is indicated
by the initial steep, positive slope), though it was emitted
in our direction. As the light traveled through space in our
direction, though it continued to recede from us, it entered
14
n=1/2 n=2/3 ( 
( n=5/6 cT
e
ac
sp
g
Time (Gy)

in
nd
pa
ex
n-

tn
no

R(t)
n=1
0

14 28 42 Non-expanding space model

Causal Radius (GLyr)

Emission Distance
2/3
R(t) t
Fig. 3. Particle horizon as a function of time for various R(t) = bt n
expansion rates. Non-expanding space and ΛCDM are shown for 
comparison. cT/e
R(t) t

Emission distance
Something important to keep in mind is, when we say that
a photon could have traveled to us from an object located at
0
3cT, we’re not saying the photon was emitted from a distance 0 T/e T
of 3cT. An object’s current distance D0 is larger (potentially Emission time
much larger) than its distance at the time of photon emission, Fig. 4. Emission distance De vs. emission time te for four different
De. This, of course, is due to spatial expansion—things are models: non-expanding, linearly expanding, matter dominated
farther apart now than they were when the light we are cur- [R(t)~ t2/3 ], and ΛCDM. T = current age of our universe = time of
rently seeing was emitted. The relationship between D0 and photon reception.
De is very simple: reception distance / emission distance = the
factor by which the universe expanded during the photon’s time regions of space that receded from us progressively more
of flight. Or slowly, reducing its recessional velocity (as is indicated by the
decreasing slope). When the light reaches its most distant
(8) location along its trajectory, De,max, its velocity with respect
to us is zero (slope is zero here). The recessional velocity of
Using this relationship, and our expression for reception dis- this region of space, De,max, must therefore have been c (this
tance, Eq. (7), we can derive the following: is true for all of the above, spatially expanding, models). After
this point, for the first time, the beam of light begins to ap-
(9) proach us. When the light finally reaches us, its velocity with
respect to us will be –c. It may seem counterintuitive for re-
Equation (9) allows us to determine how far away a given ceding light to undergo a turnaround and begin approaching
object was when the light we are seeing left it, in terms of us, but remember this light was emitted in our direction, and
emission and reception time (te and t0 ). Figure 4 is a graph only recedes from us initially because it exists in a region of
of emission distance vs. emission time for various expansion space where recessional velocity is greater than c. But as the
models: non-expanding space, constant recessional velocity, light moves through space toward us, it progresses through
matter dominated, and ΛCDM. Notice how in expanding regions of space that recede from us less and less quickly, as
space, light that has been traveling the longest (earliest emis- stated above. All light that we can see, from FTL objects, must
sion time) came from objects that were very close to us, not undergo this kind of turnaround.
far away from us. Greater light-travel time does not equate to It is sometimes stated that the expansion rate of space

THE PHYSICS TEACHER ◆ Vol. 57, February 2019 83

must be decreasing in order for receding light, emitted in our about 16 billion light-years from us.15
direction, to turn around and begin approaching us,25,30 but The conclusions above, namely that we can see FTL ob-
this is not true. As explained above, the cause for this turn- jects and that our observational limits depend on the rate
around is merely the fact that light emitted in our direction at which space expands, as well as confusion surrounding
travels through space into regions that recede from us pro- these topics, have been thoroughly discussed in the litera-
gressively more slowly. Spatial expansion does not have to be ture.15,16,20 What is unique about the discussion here is the
slowing in order for this to occur...For the linear expansion method of using, as a starting point, the idea that light can be
model, the rate never decreases (R = 0), and yet, as you can thought to accelerate away from its source in a spatially ex-
see from Fig. 4, the turnaround point for light seen today panding universe, and that Hubble’s law (a concept most stu-
occurs when the universe is T/e old. Nor is it the case, as it dents feel comfortable with) can be used to derive expressions
is sometimes stated, that accelerating expansion necessarily related to our cosmological horizons—namely, the particle
prevents us from seeing FTL objects.12,30-32 Because of light’s horizon, the cosmological event horizon, and the time-de-
“acceleration” in expanding space, even when objects recede pendent emission distance. The pedagogical value of this
from one another with an increasing rate—due to accelerat- approach is that it allows one to bypass a discussion of topics
ing spatial expansion—mutually receding FTL objects can that may present a conceptual barrier to non-experts (space-
exchange light… unless the rate of spatial acceleration is too time metrics, null geodesics, etc.) that conventionally precede
great. Which brings us to our final topic. any derivation and analysis of cosmological horizons. In ad-
dition, the above approach sheds light on subtleties about our
Cosmological event horizons observational limits that educators, and even professionals in
In contrast to constant or decreasing expansion rates, the field, sometimes get wrong.
accelerating spatial expansion creates cosmological event We began this discussion by thinking about whether or
horizons—distances beyond which we cannot exchange light. not it is possible to see objects whose Hubble recessional ve-
When spatial expansion accelerates,33 not only do regions of locity exceeds the speed of light. Attempting to answer this,
space recede from us, they accelerate away from us—the rate conceptually, led to the idea that light “accelerates” in expand-
of which depends on distance, analogous to Hubble’s law. In ing space, and that Hubble’s law provides a natural route for
non-accelerating universes our light will eventually catch up deriving expressions related to our observational limits. If
with any object, no matter how fast it recedes. In accelerating one extends these ideas a little further, subtleties regarding
universes this is not the case—our light may accelerate away the shape of our backward light cone (it’s not a cone), and
from us, but the regions of space themselves also accelerate, misconceptions about the oft-discussed horizon problem also
potentially preventing the light from ever catching up. How begin to emerge… topics for another paper. Indeed, simple
do we determine the event horizon, given some expansion questions about how light travels can sometimes lead to un-
rate? What we’re asking for is the current distance of an object expected and surprising results.
that a photon, emitted from us right now, will just catch up For an excellent and very readable introduction to the
with after an infinite amount of time (or at the end of time); science, history, and philosophy of cosmology, see Edward
so, Eq. (9) evaluated from now to tend. For an eternally ex- R. Harrison’s Cosmology: The Science of the Universe, 2nd ed.
panding universe, the event horizon is given by15,16,20 And for a more in-depth discussion about cosmological hori-
zons in the context of general relativity, see Refs. 15, 16, 18,
(10) 25, and 32.

So, for example, in accelerating universes of the form R(t) Acknowledgments
= bt n, where n > 1, as one can check, the event horizon de- The author would like to thank colleague Tristan Smith,
pends on the age of the universe at the time of emission. Spe- as well as the editors and anonymous reviewers for their
cifically, DEH = cte / (n –1). So, for n = 1.2 , say, DEH = 5cte ; helpful questions, comments, and suggestions.
and vr at this distance, by Eq. (1), is 6c (we can see FTL objects
in accelerating universes). For an exponentially expanding References
universe, of the form R(t) = beH0t , where H0 is a time-inde- 1. Georges Lemaître was the first to publish research deriving this
pendent constant (the Hubble constant is constant here), the velocity-distance relationship, commonly referred to as Hub-
event horizon distance is time independent. Specifically, DEH ble’s law.
= c/H0, and, as alluded to above and as you can check, vr at 2. “Distance” as used in this article corresponds to what a stan-
this distance is c. So, depending on how the universe expands, dard cosmology text would refer to as “proper distance.” By the
cosmological event horizons may not exist at all, they may “velocity” of a galaxy or photon, we mean the current rate at
which its proper distance is increasing, and by “acceleration”
exist but depend on the age of the universe, or they may exist
we mean the current rate at which its velocity is increasing.
and remain at a fixed distance. According to the ΛCDM mod-
3. A. Aczel, God’s Equation, (Four Walls Eight Windows, New
el, our universe currently is expanding at a rate that increases York, 1999), pp. 174, 175.
with time, and our cosmological event horizon is located

84 THE PHYSICS TEACHER ◆ Vol. 57, February 2019

4. M. Berry, Principles of Cosmology and Gravitation (IOP Pub- through space) with a Hubble recessional velocity. A discussion
lishing, Bristol, UK, 1989), p. 22. of how observational cosmological redshift supports general
5. P. C. W. Davies, The Runaway Universe (J. M. Dent & Sons Ltd., relativistic spatial expansion, as opposed to a special relativistic
London, 1978), p. 26. model, can be found in Ref. 15.
6. A. Fairall, Cosmology Revealed: Living Inside the Cosmic Egg 20. E. R. Harrison, Cosmology: The Science of the Universe, 2nd ed.
(Springer-Praxis, Chichester, UK, 2001), p. 18. (Cambridge University Press, Cambridge, 2000).
7. D. Goldsmith, The Runaway Universe: The Race to Find the Fu- 21 . Assuming a flat spatial geometry.
.
ture Cosmos (Perseus Books, Cambridge, MA, 2000), p. 52. 22. R means the rate of change of R with respect to time, or dR/dt.
8. W. J. Kaufmann, Universe, 4th ed. (W.H. Freeman and Compa- 23. The most widely accepted model for spatial expansion—
ny, New York, 1985), p. 530. ΛCDM—is of this form. Specifically, R(t) ~ sinh2/3 (t/t ) . This
9. L. M. Krauss and G. D. Starkman, “The fate of life in the uni- model accounts for the gravitational influence of cold dark
verse,” Sci. Am. 281, 36–43 (Nov. 1999). matter (CDM) and the repulsive influence of a cosmological
10. A. Liddle, An Introduction to Modern Cosmology (Wiley, Ltd., constant, .
Chichester, West Sussex, 2003), p. 21. 24. Rewrite Eq. (6) as
11. D. Overbye, “Space ripples seen as Big Bang’s smoking gun,”
The New York Times, A1, A12 (March 18, 2014). then use the quotient rule from calculus to rewrite the lefthand
12. M. Rees, Our Cosmic Habitat (Princeton University Press, side as
Princeton, NJ, 2001), pp. 166-168.
13. W. Wright and A. Wright, At the Edge of the Universe (Ellis . So, Eq. (6) becomes . Multiply both
Horwood Ltd., Chichester, West Sussex, 1989), p. 104. sides by dt and integrate from te to t0 to get
14. R. Wagoner and D. Goldsmith, Cosmic Horizons (W.H. Free-
man and Company, San Francisco, CA, 1982), p. 160.
15. Tamara M. Davis and Charles H. Lineweaver, “Expanding
confusion: Common misconceptions of cosmological hori- Set D(te) = 0—distance of photon is 0 at time of emission—and
zons and the superluminal expansion of the universe,” Pub. solve for D(t0).
Astron. Soc. Australia 21 (1), 97 (2004), http://arxiv.org/pdf/
25. W. M. Stuckey, “Can galaxies exist within our particle horizon
astro-ph/0310808v2.pdf.
with Hubble recessional velocities greater than c?” Am. J. Phys.
16. G. F. R. Ellis and T. Rothman, “Lost horizons,” Am. J. Phys. 61, 60, 142–145 (Feb. 1992).
883–893 (Oct. 1993).
26. Mathematically, the conventional approach, mentioned above,
17. P. T. Landsberg and D. A. Evans, Mathematical Cosmology: An may require fewer steps, but a grasp of several general relativis-
Introduction (Oxford University Press, Oxford, 1977). tic concepts is required.
18. A. N. Silverman, “Resolution of a cosmological paradox using 27. See Ref. 18, p. 448.
concepts from general relativity,” Am. J. Phys. 54, 1092–1096
28. Regions of space, actually. The things themselves may no lon-
(Dec. 1986).
ger exist.
19. In special relativity, velocity addition, for physical objects
29. Keeping in mind that an event horizon in between us and the
moving through space, is affected by a maximum and invariant
particle horizon will define a different kind of causal radius, a
universal speed, c, and takes the form
radius beyond which no object can receive light emitted from
us right now.
30. See Ref. 18, pp. 446, 447.
If there were no such speed limit, or if it were infinitely large 31. C. H. Lineweaver and T. M. Davis, “Misconceptions about the
( c = ∞), velocity addition would reduce to the simple Galilean Big Bang,” Sci. Am. 292, 24–33 (March 2005).
form u = u + v . Because Hubble recessional velocity is not 32. E. R. Harrison, “Hubble spheres and particle horizons,” Astro-
assumed to result from motion through space, and because the phys. J. 383, 60–65 (Dec. 1991).
space of our universe is not assumed to expand within some ..
33. The expansion of space is said to “accelerate” if R is positive.
other space—with its own maximum and invariant speed
limit—there is no need to assume a special relativistic gamma
Adam Neat is a laboratory lecturer in the Physics and Astronomy
factor limiting the rate at which spatial expansion can change
Department at Swarthmore College, in Swarthmore, PA. He has a master’s
the distance between two points (or objects), and, subsequent- degree in applied physics from Northern Arizona University.
ly, there is no need to introduce a gamma factor when adding Physics and Astronomy Department, Swarthmore College, 500
either Hubble recessional velocities together (velocities due College Ave., Swarthmore, PA 19081; aneat1@swarthmore.edu
merely to spatial expansion), or a peculiar velocity (a velocity

THE PHYSICS TEACHER ◆ Vol. 57, February 2019 85