What does a hole in a mountain mean, exactly? Daniel's been reading up on nuclear testing again, and he's landed on something that most coverage skips. We talk about nuclear tests as engineering milestones or as proliferation crises. We almost never talk about them as a communication medium.
His question is this. North Korea has turned nuclear testing into something close to an art form as a diplomatic and military signal. But the United States has also used tests at very specific moments to send very specific messages. The number of states with this capability is small. The world hopes the weapons are never used. But testing has been part of the story since the technology was born. So how necessary is testing, to the extent that word even applies? Does it have to be public? Do states test to communicate rather than to gather operational data, or do both kinds of tests coexist on different timelines?
So let's start with who actually has this capability, and what testing has meant for each of them.
Nine states. The United States, Russia, the United Kingdom, France, China, India, Pakistan, Israel, and North Korea. Israel's never declared it, but nobody serious doubts it. Only five of those are recognized nuclear weapon states under the Non-Proliferation Treaty, which tells you something about how the club has expanded since the treaty was written.
And the raw numbers on testing are staggering. Over two thousand nuclear detonations since nineteen forty-five. The United States and the Soviet Union account for the overwhelming majority of those. The United States alone conducted more than a thousand.
Two thousand holes in the ground, give or take the atmospheric shots.
Right. And that's the first thing Daniel's question forces us to do, which is separate two functions that get muddled together constantly. One, technical validation. You test because you need to know whether the design works, whether the yield matches the prediction. Two, political signalling. You test because you want someone to know something. You want them to know you can do it, or that you're angry, or that you're not backing down.
And the central tension in Daniel's question sits right there. If testing is primarily technical, why do states announce it? And if it's primarily political, why do they need to actually detonate anything?
The Comprehensive Test Ban Treaty opened for signature in nineteen ninety-six. It still hasn't entered into force, because a handful of countries with nuclear capability haven't ratified it. But the norm against testing has hardened enormously. Testing has slowed to a trickle. North Korea is the glaring exception.
So North Korea becomes the cleanest case study. Six nuclear tests between two thousand six and two thousand seventeen. And every single one of them was timed.
Not just timed. Framed. The first test in two thousand six established the baseline. They had the capability. The world could stop pretending otherwise. The two thousand nine test came after the Six Party Talks collapsed. The two thousand thirteen test came after a round of UN sanctions. Two thousand sixteen gave us two tests, one of which they claimed was a hydrogen bomb, which most analysts thought was probably a boosted fission device at the time.
And then two thousand seventeen.
Two thousand seventeen was the big one. They claimed a thermonuclear weapon. The yield estimates were in the range of a hundred to two hundred fifty kilotons, which would make it their largest by an order of magnitude. And the timing was not subtle. This came after Trump's fire and fury rhetoric. The test was the answer.
The sequence tells you exactly how the mechanism works. North Korea doesn't just test. They time tests to coincide with adversary political moments. Presidential transitions, summits, anniversaries. And then state media frames the test as a direct message. The test is the medium.
And the message works on multiple audiences simultaneously. The United States is the primary audience, obviously. But South Korea and Japan are listening too. And the domestic audience matters. The test is a demonstration that the regime is strong, that the sacrifices the population has made were worth it.
The domestic audience point is underrated. A nuclear test is the most expensive fireworks display in human history, and it's partly for the home crowd.
But here's the thing that makes North Korea's approach sophisticated. The moratorium matters as much as the tests. In twenty eighteen, when diplomacy opened up, they announced a moratorium on nuclear testing. They dismantled parts of the Punggye-ri test site, and they made sure cameras were there to film it. The absence of testing became a signalling concession.
The silence was the message.
And that's the part most analysis misses. The ability to stop testing is itself a bargaining chip. You can trade a moratorium for sanctions relief or for a summit. The test and the moratorium are both moves in the same game.
So if North Korea is the clearest case of testing as public signalling, what does the US experience tell us about the quieter side of the same coin?
The United States conducted over a thousand tests, and most of them were underground, and many of them were never announced. They were data-gathering exercises. You detonate a device at the Nevada Test Site, you measure the yield, you check the diagnostics, you file the report. The public never hears about it.
But then there are the exceptions. The ones designed to be seen.
Operation Dominic in nineteen sixty-two. This is the clearest American example of testing as deterrence messaging. The Soviet Union had broken a testing moratorium in nineteen sixty-one with a massive series of atmospheric tests. The United States responded with Dominic, which included high-altitude tests like Starfish Prime. That one detonated about four hundred kilometers up over the Pacific, and the electromagnetic pulse knocked out streetlights in Hawaii.
And the point of Dominic wasn't just to gather data on high-altitude effects, though it did that. The point was to show the Soviets that the United States could match them shot for shot.
The timing overlapped with the Cuban Missile Crisis. Dominic was already underway when the missiles were discovered in Cuba. The tests became part of the backdrop to the most dangerous moment of the Cold War. The signal was unmistakable. We are testing. We are ready. Do not miscalculate.
There's also the madman dimension. Nixon's Operation Giant Lance in nineteen sixty-nine. Not a test, but it uses the same logic. Nixon ordered nuclear-armed bombers to fly patterns near Soviet airspace for several days, deliberately visible, deliberately provocative. The idea was to convince the Soviets that Nixon was unpredictable enough to actually use nuclear weapons.
The madman theory. And it's relevant here because it shows that nuclear posturing as communication doesn't require a detonation. A test is one tool in a larger toolkit. The bombers, the missile launches, the submarine patrols. All of it is a language.
Nixon's problem was that the Soviets didn't seem to notice Giant Lance. Or if they noticed, they didn't react the way he wanted. Which raises the question of whether signalling works if the audience doesn't receive the message.
The audience problem. We'll come back to that. But first, the technical necessity question, because Daniel asked it directly. How necessary is testing?
And the answer depends on where a program is in its lifecycle.
For a new program, testing is essential. You cannot validate a warhead design on paper. The physics of a thermonuclear detonation is so complex, so dependent on the precise geometry of the implosion, the timing of the detonators, the behavior of the materials under extreme conditions, that you simply have to test. Every nuclear state tested before it had a credible arsenal. Every single one.
Including Israel.
Israel's case is murky, but the Vela incident in nineteen seventy-nine is widely believed to have been an Israeli nuclear test in the South Atlantic, possibly conducted with South African cooperation. Israel has never acknowledged it. But the pattern holds. A state with a new program tests.
For a mature program, though, the calculus changes.
The United States hasn't conducted a live nuclear test since nineteen ninety-two. That's more than thirty years. And yet the arsenal is still considered reliable. How? The Stockpile Stewardship Program. Subcritical experiments at the Nevada National Security Site, where you detonate small amounts of plutonium with conventional explosives, but you never reach critical mass. Computer simulations that model every aspect of a detonation. Surveillance of the actual weapons in the stockpile, checking for corrosion, for material degradation.
So the question of necessity shifts. For a mature program, live testing may not be technically necessary. The simulations and subcritical experiments can do the job.
That's the consensus view, with caveats. The people who run the stewardship program are confident. But there's a school of thought that says you can't fully replace empirical testing with simulation, that there are unknowns you can't model. The debate surfaces every time there's talk of resuming testing.
Which brings us to the interesting question. If simulation can replace detonation, why test at all? And the answer is that the political utility remains even when the technical necessity fades.
North Korea tested six times not because they needed six data points. They tested six times because each test was a political event. The technical data was a bonus. The real product was the headline.
So now we have the two timelines Daniel asked about. The quiet timeline, where states gather data without anyone noticing. And the public timeline, where states detonate specifically so that everyone notices.
And they coexist. The United States runs subcritical experiments on a regular basis. These are not announced. They are the quiet timeline in action. North Korea detonates a device at Punggye-ri and the whole world watches. That's the public timeline.
The quiet tests are for the engineers. The public tests are for everyone else.
And here's the knock-on effect that Daniel's question points toward. The CTBT hasn't entered into force, but the norm against testing has real teeth. When India and Pakistan tested in nineteen ninety-eight, they both faced international sanctions and diplomatic isolation. India's tests were a declaration of status. Pakistan's tests, three weeks later, were a response. Both paid a price.
And yet they tested anyway. Because the signalling value outweighed the cost.
India wanted the permanent seat at the table. The test was the price of admission. Pakistan needed to demonstrate that it could match India, that the balance of terror applied in South Asia too. The sanctions were worth it.
So the calculation is always the same. What does the test buy you, and what does it cost you? North Korea has decided repeatedly that the test buys more than it costs.
But the costs are real. North Korea's tests have accelerated US-South Korea-Japan trilateral cooperation. They've justified missile defense deployments. They've hardened attitudes in Washington and Seoul. The test signals strength, but it also provokes the coalition that is designed to contain you.
The audience problem you mentioned earlier. A test is meant to influence a specific audience. But you don't control how that audience interprets it. The message received in Washington may not be the message intended in Pyongyang.
The United States had the same problem during the Cold War. Tests that were meant to signal resolve sometimes read as provocation. Tests that were meant to reassure allies sometimes read as escalation. The sender doesn't control the receiver.
And that's the instability problem. If testing is primarily signalling, then banning testing removes a communication channel. States that can't signal through tests have to find other ways to communicate. Or they miscalculate because they can't read each other.
But the counterargument is that banning tests reduces the risk of miscalculation. If nobody can test, then nobody can misread a test as preparation for war. The CTBT is a communication protocol that says, we will not use this particular channel.
So the question is whether the channel is worth keeping. And I don't think there's a clean answer.
There's also the future question. Simulation technology keeps improving. At some point, a state might be able to conduct a virtual test, a simulated detonation, and announce it publicly for signalling purposes. No actual explosion, but the same political effect.
Would that work? Would a simulated test carry the same weight as a real one?
I don't think so. The physical act of detonation is the credibility mechanism. A simulation can be faked, exaggerated, or dismissed. A real test is a fact. The seismic signature doesn't lie. The radionuclides in the atmosphere don't lie. The physical act is the proof.
The hole in the mountain is the message.
Right. And that's what Daniel's question about public versus quiet testing really gets at. The public test is public precisely because it has to be verified. The whole point is that the other side's seismographs pick it up. The quiet test is quiet because verification doesn't matter. It's for the engineers, not for the audience.
So the two timelines serve different masters. The quiet timeline serves the technical program. The public timeline serves the political program. And a state can run both simultaneously.
The United States has been doing exactly that for thirty years. Quiet subcritical experiments for the technical program. No public tests, because the political program doesn't need them right now. The deterrent is established. The message has been sent.
North Korea runs the public timeline almost exclusively. They need the political signal more than they need the technical data, because their deterrent is still being established.
And that's the asymmetry Daniel is pointing at. The same technology, the same physics, but completely different communication strategies depending on where a state sits in the hierarchy.
Let's talk about the quiet testing timeline in more detail. The subcritical experiments.
The Nevada National Security Site, formerly the Nevada Test Site. The United States has conducted subcritical experiments there since the mid-nineteen nineties. The idea is that you can study the behavior of plutonium under shock compression without actually triggering a nuclear chain reaction.
So you're detonating conventional explosives around a small amount of plutonium, and you're measuring what happens to the plutonium.
The plutonium never reaches critical mass. There's no nuclear yield. But you get data about how the material behaves under the conditions it would experience in a real detonation. That data feeds into the computer models.
And these experiments are not announced.
They're not secret in the sense that the program is acknowledged. But individual experiments are not publicized. There's no press release. No state media coverage. The quiet timeline.
Russia and China run similar programs. So the quiet timeline is not uniquely American.
Right. And the CTBT doesn't ban subcritical experiments. The treaty bans nuclear explosions, defined as releases of nuclear energy. Subcritical experiments don't release nuclear energy. So they're legal under the treaty.
Which is one of the misconceptions Daniel's question helps clear up. The CTBT doesn't ban all testing. It bans a specific kind of testing. The quiet timeline continues.
And the verification regime for the CTBT is still being built. The International Monitoring System, the seismic stations, the radionuclide detectors, the hydroacoustic sensors. It's designed to detect nuclear explosions anywhere on the planet. But it's not complete, and the treaty isn't in force, so the system operates in a kind of limbo.
The verification question is interesting on its own. How do you know if someone is testing quietly?
Seismic monitoring is the primary tool. An underground nuclear test produces a very distinctive seismic signature. It's not the same as an earthquake. The waveform is different. The ratio of different types of waves is different. An experienced analyst can tell the difference almost immediately.
And that's where the quiet timeline and the public timeline intersect. A public test is designed to be detected. The seismic signature is the point. A quiet test is designed not to be detected, or at least not to be attributable.
Decoupling. You can detonate a device in a large underground cavity, and the cavity absorbs some of the seismic energy. The signal is muffled. The test might look like a small earthquake. This was a real concern during the CTBT negotiations.
So the quiet timeline has a spectrum. From acknowledged but unpublicized, to deliberately concealed.
And the verification community spends a lot of time thinking about this. How small can a test be and still be detected? How well can you hide it? What are the telltale signs?
This is where the technical and the political merge. The ability to detect a test is itself a political fact. If you can detect the other side's tests, you can respond to them. If you can't, you're blind.
And the other side knows this. So the public tests are a way of saying, we know you can see this, and we want you to see it. The quiet tests are a way of saying, we're doing this for ourselves, and we'd rather you didn't watch.
Let's go back to the audience problem, because I think it's the most interesting knock-on effect in Daniel's question.
The question of who a test is meant to influence.
North Korea's tests are aimed at the United States, South Korea, Japan, and the domestic audience. But the effect on each audience is different. The United States sees a threat. South Korea sees a threat and a reason to rely more on the US alliance. Japan sees a threat and a reason to build up its own capabilities. The domestic audience sees strength and defiance.
And the United States' Cold War tests were aimed at the Soviet Union and at allies. The message to the Soviet Union was, we can match you. The message to NATO allies was, we will defend you. The same test could carry both messages.
But the ally message could backfire. A test that was meant to reassure could read as escalation. A test that was meant to demonstrate resolve could read as recklessness.
And the adversary message could backfire too. North Korea's tests have accelerated trilateral cooperation between the US, South Korea, and Japan. The test that was meant to intimidate has produced a more unified opposition.
So the signalling value is not something you control. You send the message, but the audience interprets it. And the interpretation can be the opposite of what you intended.
This is the fundamental problem with signalling as a strategy. You're trying to communicate, but the communication channel is noisy. The message gets distorted.
Yet states keep doing it. Because the alternative, not signalling at all, is worse. If you have a capability and you don't demonstrate it, the adversary might not believe you have it. Credibility requires demonstration.
The demonstration problem. You can't just claim to have a nuclear weapon. You have to prove it. The test is the proof.
Which brings us back to the necessity question. Is testing necessary? For a new program, yes, because you need the technical data and you need the credibility. For a mature program, the technical necessity fades, but the political utility remains.
The political utility is not just about deterrence. It's about status. Testing is a way of saying, we are a nuclear power, we belong at the table. India's nineteen ninety-eight tests were as much about status as about security.
The status dimension is underrated. The permanent five members of the UN Security Council are all nuclear weapon states. That's not a coincidence. The nuclear club and the great power club overlap almost perfectly.
The states that tested later, India, Pakistan, North Korea, were all making a claim about their place in the international order. The test was the application.
Daniel's question about necessity has a layered answer. Technically necessary for new programs. Politically useful for all programs. And the two functions coexist on different timelines.
The quiet timeline for the technical program. The public timeline for the political program. And a state can move between them depending on what it needs at a given moment.
North Korea moved from the public timeline to a moratorium in twenty eighteen, which is a kind of quiet timeline. The absence of testing was the signal.
Then they moved back. The moratorium ended. Testing resumed, or at least the threat of testing resumed. The cycle continues.
What does the future look like? If simulation keeps improving, does the technical necessity of live testing disappear entirely?
It's already mostly disappeared for mature programs. The United States has gone thirty years without a live test. The simulations are good enough. The question is whether they're good enough forever.
The political utility? Does that ever disappear?
I don't think so. As long as nuclear weapons exist, the demonstration of capability will have political value. The test is the most unambiguous way to demonstrate.
Unless the demonstration itself becomes obsolete. If everyone knows you have the capability, and the simulations are good enough to maintain it, then the test adds nothing.
But the test adds something that the simulation can't. The physical act. The seismic signature. The fact that you actually did it. There's a difference between saying you can do something and doing it.
The difference between a claim and a demonstration.
In nuclear deterrence, the demonstration matters. Credibility is built on demonstrations. The test is the ultimate demonstration.
We're likely to see continued testing by states that need to demonstrate, and continued quiet stewardship by states that have already demonstrated enough.
North Korea needs to demonstrate. The United States has demonstrated enough. The asymmetry persists.
The CTBT sits in the middle, not in force but not irrelevant. The norm against testing is real, even if the treaty isn't binding.
The norm is the interesting part. The CTBT hasn't entered into force because a handful of states haven't ratified it. But the global expectation is that testing is abnormal. North Korea's tests are condemned precisely because they violate the norm.
The norm does the work that the treaty can't.
The norm is enforced through sanctions, through diplomatic isolation, through the costs that testing imposes. The costs are real, but so are the benefits. States weigh them.
Daniel's question about whether testing has to be public. The answer is no, it doesn't have to be. But when it is public, it's public for a reason.
The public test is a broadcast. The quiet test is a memo. Different audiences, different purposes.
The broadcast is aimed at multiple audiences simultaneously. The adversary, the allies, the domestic population, the international community. The test is a message to everyone at once.
Which is what makes it such an efficient signalling tool. One detonation, multiple messages. The cost is high, but the reach is broad.
The efficiency of the hole in the mountain.
There's a phrase for it. The test is the message.
Where does this leave us? Testing is both engineering and communication. The two functions coexist. The quiet timeline and the public timeline run in parallel. And the necessity of testing depends on where a state sits in the nuclear hierarchy.
The future is uncertain. Simulation improves. The norm strengthens. But the political utility of the test remains. The question is whether the technical utility will eventually fade to zero, and whether the political utility will be enough to sustain testing on its own.
I think the political utility will be enough for some states. North Korea will test again. The question is when and what message they're sending.
Whether the audience reads it correctly.
That's the part nobody controls.
I spent eighteen months staring at seismographs for a contractor that did verification work. This was late eighties. The job was to look at the squiggles and figure out whether they were earthquakes or detonations.
The quiet tests aren't just quiet. They're designed to be undetectable. And the public tests aren't just public. They're designed to be detected. The whole point of a public test is that the other side's seismographs pick it up.
Hilbert: I worked in that.
The laminated card.
You still have a laminated card.
Hilbert: It's in a drawer. Lists the seismic signatures. P-waves, S-waves, the ratios. An earthquake looks different from a detonation. The earthquake starts with the P-wave, then the S-wave. The detonation starts with everything at once.
Hilbert: We detected a Soviet test once. Never announced. Small yield, deep underground. The seismographs lit up and we knew. The report went up the chain and it became a diplomatic incident. The Soviets denied it, but the data was the data.
Hilbert: The card was from training. They gave us laminated cards so we could check the waveforms in the field. I kept mine. Never threw it away.
The verification side is the audience. The test is only a signal if someone's listening.
The listening is the whole point. The public test is designed for the listeners. The quiet test is designed to avoid them.
Hilbert: We were the listeners. The whole system existed so that someone would be listening. The treaty wasn't in force, but the monitoring was already running. They wanted to know if the other side was testing.
Hilbert: I predicted a North Korean test once. Years later. I was out of the business by then, but I still followed the data. The background noise patterns shifted. I told a guy I used to work with. He didn't believe me. Six weeks later, they tested.
The quiet timeline has its own listeners. That's what Daniel's question assumes. The quiet tests coexist with the public ones because the quiet tests are for the technicians and the public tests are for everyone else.
The technicians on the other side are listening to both. The quiet test that gets detected becomes a public test whether you wanted it to or not.
Hilbert: That's the risk. You think you're being quiet. You're not quiet enough. The other side's seismographs pick it up. Now you've sent a message you didn't mean to send.
Hilbert: The Soviet test we detected. They probably thought it was deep enough, small enough. They were wrong.
The quiet timeline has a failure mode. The test that was meant to be invisible becomes visible. And then the message is, we tried to hide this from you, and we failed.
Which is a different message entirely. A worse one.
Hilbert: The card's still in the drawer. I look at it sometimes. The waveforms don't change. An earthquake is an earthquake. A detonation is a detonation. The physics doesn't care about the politics.
Hilbert: That's the thing about the quiet tests. The physics doesn't care if you announce it. The data is the data. The only difference is who's looking at the readouts.
The physics as the ground truth underneath the signalling. The test is a message, but it's also just a hole in the ground with a measurable signature.
The signature doesn't lie. That's why the public test is the ultimate demonstration. You can't fake the seismic record. You can't fake the radionuclides. The physical act is the proof.
Hilbert: You can't fake it. But you can try to hide it. Sometimes it works. Sometimes it doesn't.
Hilbert: I don't miss the job. Eighteen months of staring at squiggles. But I kept the card.
The verification community is the invisible audience for the quiet timeline. They're the reason the quiet tests aren't entirely quiet. Someone is always listening.
The public tests are for everyone else. The ones who don't have seismographs but do have newspapers.
The two audiences. The technicians and the public. The quiet timeline speaks to the technicians. The public timeline speaks to the public.
Daniel's question is about how those two timelines coexist. The answer is that they coexist because they serve different purposes. The quiet tests maintain the arsenal. The public tests maintain the deterrent.
The verification community sits in between, listening to both. The quiet tests that get detected become public. The public tests that get detected are doing their job.
Hilbert: The job was to listen. That's all. Just listen.
The future question is whether the quiet timeline can eventually replace the public timeline entirely. If simulation gets good enough, and the verification community can confirm the simulations are accurate, then maybe the public test becomes unnecessary.
But the public test has a credibility that the simulation doesn't. The simulation is a claim. The test is a demonstration. And in deterrence, the demonstration matters.
The demonstration is the message. The hole in the mountain is the proof.
The proof is what Daniel's question keeps circling back to. How necessary is testing? Necessary enough that every nuclear state has done it. Necessary enough that the ones who can still do it keep doing it when they need to send a message.
The message and the data. Both at once. That's the answer.
That's the answer. The test is both. The engineering and the communication. The quiet and the public. The data and the message. They coexist because they have to. The physics and the politics are the same event.
The open question is what happens when the engineering no longer requires the event. Does the politics keep it alive on its own?
I think the politics keeps it alive for the states that still need to demonstrate. The states that have already demonstrated enough can afford to be quiet. The states that haven't will keep testing.
The rest of us keep listening.
That's the system. The listeners and the testers. The quiet and the public. The physics and the politics.
This has been My Weird Prompts. Thanks to Hilbert Flumingtop for producing.
If you found this interesting, we'd love a review. It helps other listeners find the show.
We'll be back soon.
