WEBVTT

00:00.000 --> 00:04.800
Thank you to Babel for supporting PBS.&nbsp;
Don't panic, but it's possible that an

00:04.800 --> 00:11.520
event horizon is forming right behind you right&nbsp;
now. Don't bother looking. You won't see it. It

00:11.520 --> 00:14.960
definitely won't look like the classic picture&nbsp;
of the black hole that you're probably familiar

00:14.960 --> 00:20.543
with. You won't even know that it's growing&nbsp;
to envelop you until it's way too late.

00:16.831 --> 00:22.911
for the universe's ultimate fate. Now onto the&nbsp;
episode. By now, we all have an idea of what black

00:22.911 --> 00:28.591
holes are supposed to look like. Discs of absolute&nbsp;
darkness on the sky, surrounded by gravity warped

00:28.591 --> 00:34.431
starfields, and the final electromagnetic screams&nbsp;
of stuff about to fall in. We think of the

00:34.431 --> 00:40.431
black hole as this roughly spherical surface of&nbsp;
lightlessness. And the name we give that surface

00:40.431 --> 00:46.591
is the event horizon. If I were to ask where&nbsp;
is the black hole, you might point to the event

00:46.591 --> 00:52.031
horizon. And that would be fair because a black&nbsp;
hole is really defined as something with an event

00:52.031 --> 00:58.191
horizon. The surface of no escape surrounding&nbsp;
a region so packed with mass and energy that

00:58.191 --> 01:04.911
even light becomes trapped. The event horizon is&nbsp;
as good a wear as anything for the black hole.

01:04.911 --> 01:11.471
And if I were to ask what is the black hole, it&nbsp;
seems fair to answer anything inside that horizon.

01:11.471 --> 01:15.871
Well, it turns out these answers are somewhere&nbsp;
between inadequate and just plain wrong. Even the

01:15.871 --> 01:22.031
question, when does the horizon exist isn't that&nbsp;
simple. Event horizons are actually much harder

01:22.031 --> 01:28.751
to pin down both in space and time. And when it&nbsp;
comes to black holes, a little bit of ambiguity

01:28.751 --> 01:35.711
can get you into a lot of trouble. So for safety&nbsp;
sake, let's get clear about the event horizon. If

01:35.711 --> 01:41.631
we define the event horizon as that surface from&nbsp;
which light cannot escape, we need to ask escape

01:41.631 --> 01:47.951
to where? One pretty reasonable answer would be&nbsp;
to require that in order to be considered above

01:47.951 --> 01:53.151
the event horizon, light has to be able to truly&nbsp;
escape, which means it must be able to get far

01:53.151 --> 01:58.511
enough away from the black hole that it no longer&nbsp;
is affected by its gravity. This is actually close

01:58.511 --> 02:04.831
to the formal definition. Light must be able to&nbsp;
reach infinite distance which is physics ees for

02:04.831 --> 02:10.991
very very far away. But this definition is already&nbsp;
at odds with how we tend to imagine the event

02:10.991 --> 02:18.511
horizon as the scary black disc or sphere. Let's&nbsp;
approach that disc then to investigate. We'll

02:18.511 --> 02:24.671
travel to the very near vicinity of what we think&nbsp;
of as an event horizon. Now, after collecting

02:24.671 --> 02:30.271
some Hawking radiation or whatever, we switch on&nbsp;
our ridiculously powerful rockets to move away.

02:30.271 --> 02:36.431
Because we carefully stayed above what looks to us&nbsp;
like the horizon, we're able to make some headway.

02:36.431 --> 02:41.871
But then, misfortune strikes. A giant asteroid&nbsp;
or something falls into the black hole from above

02:41.871 --> 02:48.351
us and the black hole grows in mass and the event&nbsp;
horizon expands and we're enveloped. But according

02:48.351 --> 02:53.551
to our earlier definition, if we or if light&nbsp;
cannot escape to an extreme distance, then we

02:53.551 --> 02:59.471
were already beneath the event horizon and we're&nbsp;
already doomed. Yes, something that black surface

02:59.471 --> 03:05.471
did expand to envelop us, but it wasn't the true&nbsp;
event horizon at all. So, let's dig a little

03:05.471 --> 03:10.991
deeper into the formal definition of the event&nbsp;
horizon because it completely changes the what,

03:10.991 --> 03:18.511
where, and even the when of our understanding of a&nbsp;
black hole. The textbook definition, the one that

03:18.511 --> 03:23.471
Steven Hawking and George Ellis sit down in their&nbsp;
book, The Large Scale Structure of Spacetime,

03:23.471 --> 03:30.671
is as follows. An event horizon is the boundary&nbsp;
of the causal past of future null infinity. The

03:30.671 --> 03:36.591
causal past of future null infinity. Okay, super&nbsp;
helpful. Well, we can break it down maybe a little

03:36.591 --> 03:42.351
further. We'll use a Penrose diagram. Longtime&nbsp;
viewers will remember these and it'll make

03:42.351 --> 03:48.911
things a little clearer, hopefully. Before we&nbsp;
complicate things by adding black holes, here's

03:48.911 --> 03:54.671
a Penrose diagram of an infinite flat universe.&nbsp;
Time increases upwards and one dimension of space

03:54.671 --> 03:59.951
is traced horizontally, but the contours of space&nbsp;
and time are curved so that the path of light is

03:59.951 --> 04:05.631
always 45°. And those lines are also increasingly&nbsp;
squished together towards the edges so that

04:05.631 --> 04:11.311
infinite space fits in this finite map. The points&nbsp;
at the top and bottom represent our infinite

04:11.311 --> 04:17.871
future and past. And the left and right points are&nbsp;
infinite spatial boundaries. But it's the diagonal

04:17.871 --> 04:24.031
edges of the diagram that we need to pay attention&nbsp;
to. These are the null infinities that Hawking and

04:24.031 --> 04:30.351
Ellis are talking about. In particular, the&nbsp;
future null infinity is where light rays end

04:30.351 --> 04:35.151
in the infinite future. And the past null infinity&nbsp;
is where they appear to come from if we trace them

04:35.151 --> 04:41.471
back in time forever. Now we can add a black&nbsp;
hole to the Penrose diagram like this. This

04:41.471 --> 04:48.031
is an eternal black hole. It was always there and&nbsp;
always will be there. It's just sitting out there

04:48.031 --> 04:53.551
to our left and any motion in that direction will&nbsp;
eventually reach it. And that means we can replace

04:53.551 --> 05:00.591
one future null infinity by the event horizon.&nbsp;
Below the event horizon, space and time grid lines

05:00.591 --> 05:06.751
actually switch place. Moving up on this diagram&nbsp;
is equivalent to falling deeper in the black hole

05:06.751 --> 05:13.151
towards the inevitable destination of the black&nbsp;
hole center. the singularity which occupies all

05:13.151 --> 05:18.511
future times inside the black hole. We have&nbsp;
previous episodes that explain all of this in

05:18.511 --> 05:23.951
more detail. With this sort of diagram, we can see&nbsp;
that nothing can escape the black hole. In order

05:23.951 --> 05:29.231
to escape, something would need to take a path&nbsp;
shallower than 45° and that means going faster

05:29.231 --> 05:36.031
than light. Even light gets stuck forever at the&nbsp;
event horizon if it starts there and inevitably

05:36.031 --> 05:41.311
at the singularity even if it was trying to&nbsp;
move outwards. Hawking and Ellis's definition

05:41.311 --> 05:46.831
is clearer in the context of this diagram. They&nbsp;
define an event horizon as the boundary of the

05:46.831 --> 05:52.911
causal past of future null infinity. This is&nbsp;
the future null infinity and its causal past is

05:52.911 --> 05:59.551
anything that can get to it traveling no greater&nbsp;
than light speed. That means the entire universe

05:59.551 --> 06:06.271
except the region below the event horizon. Nothing&nbsp;
behind that boundary can affect future infinity or

06:06.271 --> 06:12.751
indeed anything else in the universe that isn't&nbsp;
itself behind the same event horizon. Okay, so

06:12.751 --> 06:18.991
it feels like we're saying the same thing that we&nbsp;
already knew but with more words and more lines.

06:18.991 --> 06:24.831
But crucially, we we've now been precise enough&nbsp;
about black holes as a causal structure. There's

06:24.831 --> 06:29.631
something that determines what events are allowed&nbsp;
and forbidden in the universe. In doing this,

06:29.631 --> 06:35.711
we remove a lot of ambiguity that we ran into&nbsp;
earlier. But we also discover something really

06:35.711 --> 06:42.271
strange. The event horizon definition is no longer&nbsp;
local. It doesn't refer to a single location in

06:42.271 --> 06:48.031
space or point in time. It's the boundary below&nbsp;
which neither you nor light rays can escape to a

06:48.031 --> 06:52.831
great distance, aka reach future null infinity,&nbsp;
even if the black hole that eventually traps

06:52.831 --> 06:58.351
you hasn't started forming yet. So the event&nbsp;
horizon is a property of the entire spaceime

06:58.351 --> 07:03.071
reaching all the way into infinite future. It's&nbsp;
not just a statement about what is possible now,

07:03.071 --> 07:08.591
but what will be possible based on everything that&nbsp;
will happen to this patch of spaceime. Physicists

07:08.591 --> 07:14.431
sometimes describe the event horizon as being&nbsp;
teological from the ancient Greek term telos

07:14.431 --> 07:20.111
meaning final purpose. It tells us about the final&nbsp;
state of the universe, the paths that make it out

07:20.111 --> 07:25.631
to infinity and those that eventually end up stuck&nbsp;
inside the black hole. That's what I mean when I

07:25.631 --> 07:31.311
say that an event horizon might have just passed&nbsp;
you right now and you wouldn't even realize. Let's

07:31.311 --> 07:38.031
do another thought experiment to see why. We'll&nbsp;
start back on Earth. Relatively flat space free of

07:38.031 --> 07:43.791
black holes. However, for unfortunate reasons that&nbsp;
we don't need to get into, but may involve vogons,

07:43.791 --> 07:50.271
there's a spherical shell of radiation of light&nbsp;
collapsing towards us. This is an example of the

07:50.271 --> 07:58.431
Va spacetime devised by Indian physicist Prahalad&nbsp;
Vaja as a toy model to explore spaceimes and event

07:58.431 --> 08:04.271
horizons of a collapsing star without all of&nbsp;
the messy matter. In this thought experiment,

08:04.271 --> 08:09.711
it feels like there's no event horizon as the&nbsp;
shell of radiation converges on us. We don't

08:09.711 --> 08:16.031
even know that it's coming because no signal can&nbsp;
travel faster than the radiation zone collapse.

08:16.031 --> 08:20.991
It's ultimately going to form a black hole with&nbsp;
a radius governed by the amount of energy in that

08:20.991 --> 08:26.591
radiation. But according to our formal definition,&nbsp;
the event horizon forms long before the black

08:26.591 --> 08:32.191
hole. In fact, it grows from the center of the&nbsp;
collapsing sphere at the speed of light, reaching

08:32.191 --> 08:38.111
its maximum size when it intercepts the infalling&nbsp;
radiation. That's when the black hole truly forms,

08:38.111 --> 08:43.311
but the horizon was already there. To visualize&nbsp;
this, let's say that you're at the center of the

08:43.311 --> 08:48.751
collapsing shell and there's a lastditch effort&nbsp;
to escape. You shoot a lighteed signal out into

08:48.751 --> 08:54.191
space to see if any passing spaceships will&nbsp;
beam you up. If your signal makes it beyond

08:54.191 --> 09:00.031
the ultimate size of the event horizon before&nbsp;
the incoming radiation crosses that same point,

09:00.031 --> 09:07.151
then it escapes and maybe you hit your way out of&nbsp;
there. But if not, then it and you remain stuck

09:07.151 --> 09:14.351
behind the horizon. If you send your signal too&nbsp;
late, then it never reaches future null infinity.

09:14.351 --> 09:20.751
So by our definition, it was always behind the&nbsp;
true event horizon, trailing it as it expands

09:20.751 --> 09:27.471
to become a black hole. Let's see what the Penrose&nbsp;
diagram looks like for a Vega spacetime. It starts

09:27.471 --> 09:34.111
like our black hole-free diagram, which we'll&nbsp;
cut in half because who cares about this entire

09:34.111 --> 09:39.711
side of the universe? The sphere of collapsing&nbsp;
radiation looks like this single line in our one

09:39.711 --> 09:46.271
dimension of space. It eventually reaches us, but&nbsp;
the black hole it forms comes fully into existence

09:46.271 --> 09:53.231
when it collapses below that certain radius. The&nbsp;
event horizon is then a diagonal line upwards from

09:53.231 --> 09:59.631
that point. Just as in our black hole containing&nbsp;
Penrose diagram, the spaceime within that event

09:59.631 --> 10:07.231
horizon is terminated by a singularity. Now, it's&nbsp;
easy to see which light rays escape the forming

10:07.231 --> 10:13.391
event horizon and which don't. Ones that we send&nbsp;
out early enough to reach future null infinity

10:13.391 --> 10:18.191
are the ones that escape. Those that depart too&nbsp;
late fail to pass the involing radiation before

10:18.191 --> 10:24.591
the black hole forms. But these rays were doomed&nbsp;
from the very start. And so we need to trace the

10:24.591 --> 10:30.591
true event horizon backwards to encompass this&nbsp;
entire region. We have a part of the universe

10:30.591 --> 10:36.911
that is trapped even though it doesn't know that&nbsp;
it's in a black hole. So yeah, a nasonent event

10:36.911 --> 10:42.671
horizon might be forming behind your couch as&nbsp;
we speak. You might want to check that out.

10:42.671 --> 10:48.191
Although you'll have to wait until the end of the&nbsp;
universe before you can be absolutely certain. A

10:48.191 --> 10:54.911
secret event horizon could be due to the collapse&nbsp;
of something huge, a giant region of the galaxy

10:54.911 --> 11:01.631
or even the universe billions of years from now.&nbsp;
All of this is a weird way to think about event

11:01.631 --> 11:08.111
horizons, but does anything really change? Who&nbsp;
cares if we can't find the horizon? Well, the

11:08.111 --> 11:13.711
most obvious problem is that an event horizon is&nbsp;
used to define the boundary of a black hole. So,

11:13.711 --> 11:21.471
it makes it hard to talk about where a black hole&nbsp;
is. It contradicts our astrophysical intuition.

11:21.471 --> 11:27.471
We know that there is a black hole at the center&nbsp;
of our galaxy called Sagittarius A star and that

11:27.471 --> 11:33.471
stars will get eaten if they get within a certain&nbsp;
distance. But this local description feels at odds

11:33.471 --> 11:40.031
with the teological definition. This is even more&nbsp;
apparent when we try to simulate the merges of

11:40.031 --> 11:46.031
black holes on a computer. Because Einstein's&nbsp;
equations are difficult to work with in these

11:46.031 --> 11:52.111
highly dynamical regimes, we rely on simulations&nbsp;
to tell us what the spaceime will look like, we

11:52.111 --> 11:57.391
set up Einstein's equations to describe two black&nbsp;
holes in spiraling, then evolve the simulation

11:57.391 --> 12:03.151
forward in time until the black holes merge, and&nbsp;
we're left with a single black hole. But where

12:03.151 --> 12:10.591
and with what shape are the event horizons of the&nbsp;
merging black holes and of the final result? If

12:10.591 --> 12:16.991
a horizon is teological, then it's defined across&nbsp;
the entirety of the simulation to future infinity.

12:16.991 --> 12:22.591
We could get a single time slice out to see how&nbsp;
spacetime has evolved, but it won't tell us where

12:22.591 --> 12:29.871
the horizon is. We could possibly trace light rays&nbsp;
back in time from the end of the simulation and

12:29.871 --> 12:34.991
look for the boundary of a region from which no&nbsp;
light gets out alive. And that's very much in the

12:34.991 --> 12:40.351
spirit of Hawking and Ellis's definition. But this&nbsp;
is very awkward as a way to find black holes and

12:40.351 --> 12:47.551
not necessarily very instructive if we care about&nbsp;
the local shape of spaceime during the simulation.

12:47.551 --> 12:53.551
And besides, we still want to know what to call&nbsp;
that black surface that we think we're escaping

12:53.551 --> 12:57.631
but are really doomed because the real event&nbsp;
horizon is already above us. In the special

12:57.631 --> 13:03.151
case of a black hole that's done forming and will&nbsp;
never absorb new stuff, that black surface really

13:03.151 --> 13:09.151
is the event horizon in the technical sense that&nbsp;
we just defined. But in the real universe where

13:09.151 --> 13:15.151
black holes tend to keep growing, we need another&nbsp;
name for the current snapshot of the black holes

13:15.151 --> 13:22.751
boundary. And that name is the apparent horizon&nbsp;
where the formal event horizon represents the

13:22.751 --> 13:28.591
global inescapable boundary integrated across all&nbsp;
future time. The apparent horizon is a more local

13:28.591 --> 13:35.951
beast. It's the locally inescapable boundary,&nbsp;
the thing that you can't exit even right now,

13:35.951 --> 13:41.471
regardless of whether you can make an ultimate&nbsp;
escape. The formal definition of the apparent

13:41.471 --> 13:47.951
horizon is also in terms of what light rays can&nbsp;
do. But now it's not whether they can reach future

13:47.951 --> 13:53.871
null infinity. It's whether they can make any&nbsp;
progress whatsoever in the outward direction.

13:53.871 --> 13:58.191
Technically, we say that the apparent horizon&nbsp;
is the outermost surface from which all outward-

13:58.191 --> 14:05.631
facing null geodysics are converging inwards.&nbsp;
Having a local definition of the horizon is nice,

14:05.631 --> 14:11.231
especially because it lets us understand what&nbsp;
things look like if we're not at null infinity

14:11.231 --> 14:17.391
or otherwise really far away. When we see these&nbsp;
awesome simulations of merging black holes,

14:17.391 --> 14:23.871
what we're seeing is the evolution of the apparent&nbsp;
horizon. So why bother with the first definition

14:23.871 --> 14:30.271
of the event horizon at all? Well, it's because&nbsp;
the apparent horizon is relative. Every observer

14:30.271 --> 14:36.191
will agree on the definition of the true event&nbsp;
horizon, the ultimately inescapable region. But

14:36.191 --> 14:42.591
the definition of the apparent horizon varies not&nbsp;
just over time, but also with frame of reference.

14:42.591 --> 14:49.231
In relativity, your sense of now varies with your&nbsp;
velocity and the curvature of spaceime. Each sense

14:49.231 --> 14:55.631
of the present is a 3D spatial slice out of&nbsp;
the four-dimensional spaceime. And that slicing

14:55.631 --> 15:01.791
changes with perspective. The apparent horizon&nbsp;
will also change between these slicings. And it's

15:01.791 --> 15:08.111
even possible to find slicings or definitions of a&nbsp;
specially extended now that include a black hole,

15:08.111 --> 15:14.431
but in which the apparent horizon disappears.&nbsp;
But that's not true of the true event horizon.

15:14.431 --> 15:20.511
If you're behind that, it doesn't matter how&nbsp;
you slice spacetime, you are stuck. In fact,

15:20.511 --> 15:27.951
every possible apparent horizon is guaranteed to&nbsp;
always lie inside or on the actual event horizon.

15:27.951 --> 15:32.591
In this sense, the apparent horizon gives a&nbsp;
conservative estimate about the extent of the

15:32.591 --> 15:39.071
black hole. It also gives a much more useful way&nbsp;
to study the behavior of dynamical black holes.

15:39.071 --> 15:46.431
black holes that are currently feeding like in&nbsp;
quazars or in merging black holes. In that case,

15:46.431 --> 15:51.551
the apparent horizon does something quite&nbsp;
unexpected. You have these apparent horizons for

15:51.551 --> 15:57.231
the inspiraling black holes that warp as they get&nbsp;
close, but the apparent horizon of the final isn't

15:57.231 --> 16:03.711
just a smooth merging of the two. That larger&nbsp;
joint apparent horizon forms only in the instant

16:03.711 --> 16:09.551
that the event horizons merge. So, the apparent&nbsp;
horizon is far from a substitute for the event

16:09.551 --> 16:16.191
horizon. It's a genuinely different object with&nbsp;
its own interesting properties. So, what, where,

16:16.191 --> 16:22.991
and when is an event horizon? And is there one&nbsp;
forming behind you right now? Well, if you care

16:22.991 --> 16:28.431
about what's happening close to a black hole at&nbsp;
a particular time, then you need to get creative.

16:28.431 --> 16:34.911
You need to choose reference frames and ask where&nbsp;
for that observer light appears to turn inwards at

16:34.911 --> 16:40.191
the apparent horizon. But if you want the most&nbsp;
clear definition, the definition that matters

16:40.191 --> 16:45.871
if you want to escape a black hole, then the true&nbsp;
event horizon is not a property of local space or

16:45.871 --> 16:51.471
immediate time. It's a thing defined by the past&nbsp;
and the future and exists as a causal boundary

16:51.471 --> 16:58.271
cutting its way through spaceime. Thank you to&nbsp;
Bubble for supporting PBS. If you spend any time

16:51.471 --> 16:56.431
in learning a new language and sharing that with&nbsp;
five other people in your life, check out the link

16:51.471 --> 17:12.591
in the description for the Babel Group plan.&nbsp;
You can get up to 55% off your subscription.
