Red Cloud at Dawn: Truman, Stalin, and the End of the Atomic Monopoly

• Introduction
• Remarks
• Questions and Answers

JOANNE MYERS: Good afternoon. I'm Joanne Myers, Director of Public Affairs Programs, and on behalf of the Carnegie Council, I would like to wish you all a very Happy New Year and thank you for joining us.

Our guest this afternoon is Michael Gordin. Professor Gordin teaches at Princeton and specializes in the history of the modern physical sciences and Russian history. He will be discussing his most recent publication, Red Cloud at Dawn: Truman, Stalin, and the End of the Atomic Monopoly. In this book he tells the dynamic story of nuclear proliferation and the end of the American monopoly of the bomb, as viewed through the lens of espionage.

Professor Gordin has published articles on a variety of topics, including the introduction of science into Russia in the early 18th century and the history of biological warfare in the late Soviet period. He has also worked extensively on the early history of nuclear weapons. Additionally, he is the author of Five Days in August: How World War II Became a Nuclear War, which is a widely acclaimed history of the atomic bombings of Japan during World War II. Professor Gordin is also the co-editor of the four-volume Routledge The History of Modern Physical Science and Intelligentsia Science: The Russian Century, 1860-1960.

The world's first atomic bomb was detonated by the Americans on July 16, 1945, in Alamogordo, New Mexico. This explosion was followed by several others, including those that fell on Hiroshima and Nagasaki. But it wasn't until the Soviets developed a bomb and tested it on August 29, 1949, that the arms race began in earnest. This event is one that would ultimately lead to nuclear proliferation beyond the two superpowers. Known as "First Lightning" to the Russians and to Americans as" Joe 1," which was a cheeky reference to Joe Stalin, this bomb had roughly the equivalent in yield to the atomic bomb that the United States had dropped on Nagasaki four years earlier.

The successful Soviet test came as a profound shock to the West, who believed that the Soviet Union was at least several years away from being able to detonate a nuclear device. Nevertheless, the fallout from the red mushroom cloud which rose over the desert of Kazakhstan is still in many ways with us today.

In Red Cloud at Dawn, Professor Gordin shares a fascinating behind-the-scenes account of espionage, secrecy, deception, political brinkmanship, and technical innovation. He challenges conventional technology-centered nuclear histories by looking at the prominent role that atomic intelligence and other forms of information play in the uncertainties of nuclear arms development and political decision making.

Our speaker had access to newly opened archives. Therefore, he was able to focus on the extraordinary story of "First Lightning" to provide a fresh understanding of the origins of the nuclear arms race, as well as the all-too-urgent problem of proliferation.

In a world still fraught with nuclear tensions, secrecy, and deception, any attempt to understand the past can be useful in facing the challenges of today. With many countries exhibiting nuclear ambitions, how we arrived at this place in time has implications for us all.

With this in mind, please join me in giving a very warm welcome to our guest this afternoon, Michael Gordin.

MICHAEL GORDIN: Thank you, Joanne. Thank you all for coming. Thank you for the invitation to speak to such an august audience of people who know quite a bit about this material, too.

I'm going to give you a historian's pass at what this all looks like, which I believe has quite a lot to say about how one should think about the present and the future. But primarily it's going to be about the past.
Red Cloud at Dawn is an international history of the atomic monopoly. Let me start by breaking down what I mean by that.

The atomic monopoly is the period from August 1945 to August 1949. That period of time is the only period in which only one nation has had the power to deliver nuclear bombs in wrath or explode them in any capacity whatsoever. It's a period bounded by explosions. For our purposes, the first explosion is the destruction of the city of Hiroshima on August 6, 1945. The other explosion is on August 29, 1949, at 7:00 a.m. local time, which is 4:00 a.m. Moscow time, at a place called Semipalatinsk-21, which is in the steppes of Kazakhstan.

That test, the Soviet nuclear test, is the ninth nuclear explosion on the earth's surface and it is the first nuclear explosion not under American auspices, in which respect I think it deserves to be called the first proliferation—that is, the first detonation of a nuclear bomb outside of a regime in which everybody had already gotten used to the idea of there being a nuclear power, although not necessarily happy about it.
But this is not really a story about atomic bombs. It's fundamentally a story about intelligence.

The Americans' quest to maintain their monopoly on this particular weapon and the Soviets' quest to break that monopoly are fundamentally bounded by intelligence. What do I mean by intelligence? I'm sure a lot of you have your own definitions. I would call it the dedicated gathering and interpretation of information about international matters.

That's a tricky business, since every time you are trying to gather information in order to make decisions about how to interact with an opponent, that opponent is trying very hard to deny you precisely the information you seek. Every time a party—say, the Americans—decides to divulge a particular fact, that raises a host of questions: Is this a reliable piece of information? Why tell us this specific piece of information and not another? What did you leave out? Why tell us now?

Those sets of questions ramify throughout the early Cold War.

The story in Red Cloud at Dawn is also fundamentally international. The Soviet quest for a nuclear bomb of its own happened under the auspices of a world in which Americans had a nuclear bomb and, it bears saying, had used this bomb twice. I'll get back to that. So it's impossible to tell the Soviet story without constantly reflecting upon what the Americans are doing. It's impossible to tell it internally to the Soviet Union. At the same time, it's also impossible to tell the American story of the monopoly without focusing at great length on what the Soviets are doing, or what the Americans think the Soviets are doing, because literally every decision that is made in the monopoly and for a good time after by the Americans is shadowed by what they think the Soviets might be up to and what its implications would be. So the whole picture can only be grasped by looking at both of them side by side.

Given that in this period the Americans are the only ones who have this particular weapon, they generate most of the documents, most of the traces, and most of the concrete decisions. So the story is about two-thirds American, one-third Soviet. I don't think there is any way to change that balance.

But fundamentally it's a story about the international interaction between the two, for nuclear weapons have been international since the moment of their first use, and they have been objects of international negotiation and contestation.

There are a great many highly abstract problems—what philosophers would call epistemological problems—that become incredibly concrete when looked at through this lens. For example, every time the Americans and the Soviets interacted, there was this constant problem of how one party knows what the other party knows about what the first party knows about what the second party knows, and so on down the line. This is called the problem of common knowledge by philosophers, and it's a very irreducibly serious problem in terms of how you are going to decide what to divulge, because you might give away the fact that you don't think the other side knows this particular fact, which in itself is an important piece of information.

In addition to the problem of common knowledge, there are also three other highly related problems:

• • The first is, how do you get information from a remote entity, whether that is the inside of the Soviet Union, what Stalin's intentions are—that is, inside his mind—or underground uranium deposits? How does one access a remote entity?

• • The second is, how do you make a concrete prediction about a future event when all of your variables are uncertain?

• The third is, how do you take that uncertain prediction and turn it into some kind of concrete policy decision? In the end, you can't just say, "Oh, sorry, my information is terrible. I can't give you a prediction on when the Soviets are going to have an atomic bomb," because that fact changes almost everything about your defense and diplomatic posture, and you have to make some decision, even though the Americans and the Soviets both knew that all their information was of terrible quality.

Each chapter of the book follows a different aspect of the problem with intelligence:

• • How much can you give away about the atomic bomb without giving away the whole secret? How much can you reveal without revealing everything?

• • Second is, how do you infer from limited data a prediction about future events?

• • Then, how do you actually make a decision when all of your information is from dubious sources, like, for example, jittery scientist spies? How do you then take that uncertain information and make it into something as concrete and real as an atomic bomb and then explode it?

• • Then, how do you find out that something has happened in a remote entity like that, and how do you persuade political actors that, in fact, your evidence is reliable and that you can act on it?

• Finally, how do you make actions in a world where every action generates counter reactions and a possible misstep could lead to disaster?

Those are the issues raised by Red Cloud at Dawn. I'm going to give you four concrete historical instances of what happens in the book, from different parts of it, and then conclude with some reflections about now.

The first is the question posed in the title of Chapter 2 of the book: How much time do we have? That is, how did American actors—political pundits, high-level scientists, politicians, elected and not, military officers—how did they reason about the problem of how long the atomic monopoly was going to last? The predictions range quite widely, and they range widely both in public documents and in classified documents which have since been declassified. The predictions generally range from two years to 20 years to 40 years to never. That's the span of what people think the monopoly's duration is going to be.

Part of the reason for this is that there are basically two data points from which you must infer. The Americans managed to make an atomic bomb in three and a half years, under wartime conditions. The Germans managed to make an atomic bomb in never. So the Soviet Union is most likely to fall somewhere between 3.5 years and never.

But that's an extremely weak prediction. So what people had to do—and I'll give you some concrete examples in a second—is evaluate the information you have and try to figure out where on that sliding scale the Soviet Union actually happens to be.

Part of the problem is the intelligence infrastructure of the United States in the monopoly is extremely weak and fragmented. In September 1945, Truman, by executive order, abolishes the Office of Strategic Services, which was the then-standing intelligence agency. There has been a longstanding tradition of distrust in the United States of peacetime intelligence-gathering agencies. Truman didn't like the OSS. Most members of Congress, of which he had recently been one, didn't. He was glad to see it go. That's September 1945.

The Central Intelligence Agency is established in September 1947. That's two years without a centralized intelligence-gathering infrastructure. It doesn't mean no one is gathering information. Obviously, the Navy, the soon-to-become Air Force—the Army Air Forces, I should say—the Army, various other groups, such as the State Department and so on, have their own intelligence groups that gather information. But there is no correlation of them.

So in the end, when the CIA is formed, it's a fledgling organization which doesn't have the capacity to process all this information and produce predictions right away, which is not surprising. But that's two years without concrete advice about a matter like this.

So what we have is basically extremely limited information. There are no human spies from the West in the Soviet Union in this period—zero, none at all. There are attempts at infiltration, but they don't work. So what you have are public press reports, some stolen documents, defector testimony from émigrés and displaced persons, some reports that are crossing the border in the German zone, and then you have speculations based on information you have, private information you have or classified information you have. You have to run that limited information through a series of filters.

The first assumption filter is, how hard is it to build an atomic bomb? If you are General Leslie Groves, the wartime director of the Manhattan Project, the answer to that is, quite difficult. It's extremely hard to build an atomic bomb. It will take the Soviets, who have a much impoverished engineering infrastructure, an extremely long time to do. He eventually hardens down to 20 years. That's his estimate. So a nuclear Soviet Union will be a problem for the grandchildren of America, not Americans of that day.

If you are the theoretical physicists who were staffing Los Alamos, the answer to how hard it is to build an atomic bomb is, not very. All the physics is basically obvious; it just requires some small engineering, which they have a low opinion of. (Groves happened to be an engineer.) And they think it's therefore not very hard at all. Nobel Laureate Irving Langmuir, for example, makes predictions before Congress in the fall of 1945 of two and a half years.

So depending on your assumption about the difficulty of building a bomb, you come up with a different number.

Another assumption is, how smart do you think the Soviets are? Military officials and diplomats tend to think the answer is, not smart at all. That's partially based on private information of whom they interacted with during the war. They have a sense that these people are lunkheads who have no physical training whatsoever—therefore, high number.

The physicists have a great deal of experience with people like Lev Landau, Vladimir Fock, Igor Tamm, who are world-class, extremely gifted physicists. They therefore think that the infrastructure in the Soviet Union is just fine. They are quite smart. It won't take very long at all. Again, the numbers tend to congeal around five years.

Third assumption: When did they start? You can say five years. That's a perfectly good estimate. But from when? No one has any idea, in this period, when the Soviet Union started making a bomb. The Soviet Union, in fact, started making a bomb three different times. They started making one in 1940, they started making it again in 1943, and they started making it in earnest in 1945. But no one knew that at the time.

Fourth assumption: How rare is uranium? Again, no one knows. General Groves managed late in the war, in his Murray Hill Area Project—all Manhattan Project codenames have some New York geography involved in the coding—in the Murray Hill Area Project, Groves manages to corner 97 percent of the world's known uranium deposits, which is quite an impressive feat. Most of that is from the Belgian Congo or from Brazil. There is a small amount of Swedish uranium that he can't get his hands on, and some stuff from Czechoslovakia.

But that's what is known. Uranium, until December 1938, January 1939, when the discovery of fission is announced, is a not-very-widely-prospected mineral. It has very few uses. Its major use is to make very bright yellow glazes for pottery and ceramics. It's not widely in demand. No one quite knows how much there is. There are very few Soviet sources of uranium, as far as they know. But the Soviet Union occupies one-sixth of the earth's land surface. Depending on whether you think uranium only exists in pockets in Brazil and the Belgian Congo or it is very widely distributed, you get a different estimate for how hard it will be for the Soviets to get a bomb. In fact, uranium is the bottleneck for the Soviet project.

The final assumption is, how much espionage has leaked? The Americans know from 1943, early 1944, that espionage has been dribbling out of the project. They don't know how much. They don't know what has specifically dribbled out. They don't decode the Venona documents until, partially, in 1949. Even that takes much longer to act upon. So they have no idea what's missing. Depending on what's missing, you could shorten the amount of time for the Soviets to get a bomb dramatically.

Based on how you answer all these, you get a different number. By 1949, people tend to say five to ten years. But that's just a compromise number that no one has any particular reason for justifying. You can't ever persuade anyone that you actually have a better number than they do. Part of the reason General Groves thinks 20 years is something he can't say, which is that he has a monopoly on the world's uranium. But that's classified, so he can't tell anyone that that's why he thinks 20 years. He has to say something else about how hard it is to build an atomic bomb.

Likewise, other people have private information or classified information that members of Congress don't have security clearances to see.

So you can never persuade anybody else that your number is better than theirs because you have this overlapping security-clearance problem. That's a perpetual problem throughout this atomic monopoly period and beyond. Yet that doesn't mean that you can just say, "We don't have any number. We don't know what's going on," because, depending on whether you think it's five years or 20 years, that radically changes your expenditures on military equipment. Do you just load up on atomic bombs in the small time period or do you try to build a conventional army? That's much, much more expensive.

Those decisions are hotly debated, especially after 1946, when Truman loses control of the houses of Congress, and a Republican majority comes in and makes it quite difficult to negotiate a particular policy outcome.

So that's case one.

Case two is the view from the Soviet Union. One of the things I try to do in the book is give a sense of what the Soviet atomic project looks like from the inside. Soviet scientists didn't know in 1945 that in four years they would have an atomic bomb. They knew a couple of things. They knew that the Americans had this weapon, that the Americans had used it twice, and used it on a non-nuclear-armed opponent. I'll get to the issue of deterrence in a second.

They also knew something which the Americans had not known in the beginning of World War II, which is that the problem is soluble, that, in fact, it is possible to build an atomic bomb. Hiroshima and Nagasaki dispel that question, which is an active question until about 1943 and the American project.

So they set about building a very large infrastructure, which is directed by Lavrentiy Beria, who is one of Stalin's right-hand men and is, although not officially, actually in charge of Soviet internal and foreign intelligence. They gather lots of information. This raises the issue of espionage, which is the central issue that is talked about by historians and policymakers.

The issue of espionage is actually two different questions. The first is, how important was espionage as a source of information? That is, among the many sources of information, how important is espionage? The second is, how was that espionage information used within the Soviet atomic bomb project?

First, the question of how important it was. It's important. It's not as important as people tend to think. And that's not because American information isn't important; it's because a huge amount of information is released in open-domain/open-source material. The Smyth Report, which is the official history of the Manhattan Project, written by the chair of Princeton's Physics Department, Henry DeWolf Smyth, is released on August 12, 1945, before the war is even over. It's very carefully vetted so that no classified information, according to Groves, is released.

But it releases a lot of information, just nothing that—the standard is, nothing that a well-equipped laboratory could not figure out within one to two years. So a lot of stuff is released. Based on what's released and what's specifically omitted, they can tell a lot of information about what's going on. The same is true for congressional testimony, interviews, newspaper articles, and so on. A lot of that information is extremely important for the development of the Soviet atomic bomb.

That said, there are about 1,000 pages of very pertinent spy information—what I like to call spy data, to make an analogy to scientific data—that is gathered from the Manhattan Project in various places, principally from Los Alamos. One particular piece of information, which they get twice—from Klaus Fuchs, who is a German émigré, British subject, working at Los Alamos, and Ted Hall, who is the youngest of the physicists working at the project—is crucially important, which is that you cannot make a gun-style bomb with plutonium. A gun-style bomb is when you take two lumps of subcritical uranium and launch them into each other down a tube. This is what the "Little Boy" bomb is structured like. You can make it out of purified uranium. The problem is, it's very hard to purify that much uranium, which you need. You need something well above 80 percent, towards 95 percent, to make that happen—that is, purified uranium of a certain isotope, which we can talk about later if you would like.

The problem is, if you do it with plutonium, it pre-detonates. Plutonium is much easier to get your hands on, because you can make it out of the dross uranium, which makes up over 99 percent of naturally occurring uranium. The only problem with that is that you cannot use the simple gun mechanism. You have to put lots of subcritical masses around a sphere and implode it into a hard core.

What Fuchs and Hall tell the Soviets is, if you want to make a bomb with plutonium, you must design an implosion mechanism. They don't have any scales. They don't have any formula. They show them a basic idea of implosion and say, "You guys are going to have to work it out on your own."

That piece of information is crucial. The first Soviet bomb is, in fact, a plutonium implosion bomb, which is what the Nagasaki bomb, "Fat Man," also was.

So it's important in certain specified areas, but most of the information they need, such as how many plants you need to build, what the basic order of processes is, and a basic manual of what people who are workers in the factory need to know—that stuff is publicly available. The Soviets take it and translate it and distribute it very widely.

The spy information, less so. This is the second question. How was it used in the project? Very few people have access to the spy information. Basically, Beria controls it very strictly. His scientific adviser, Igor Kurchatov, also controls it very strictly. Only the heads of very large laboratories and their deputies are allowed to look at it. You have to sign and countersign an affidavit saying, "I have, in fact, looked at this information and agree not to spread it." Beria is a very scary person. No one particularly wants to spread it beyond him.

So that information is extremely classified within the Soviet context. There was tremendous shock among Soviet physicists in the late 1980s, when revelations about the extent of spy information was revealed. They just assumed that they had built it on their own and that allegations of espionage were an American propaganda ploy, because they genuinely didn't know.

So, first of all, there is a problem of distribution.

Second, there is a problem of completeness. Spy information tends to be haphazard. First of all, it's given to you by people you don't trust. They are called spies. It's kind of an issue. Second, the information you get tends to be incomplete. It's a couple of papers drawn out of some files. It's randomly gathered. It's often out of date, quite severely so. Not a single piece of information the Soviets get through these documents has a scale on it, for example, so they don't know how big these factories have to be.

There is an additional issue of not trusting the information in the first place. You can trust your spy, but you could believe it's all misinformation. So every single piece of spy data they have is checked and double-checked and triple-checked by Beria. They spend a lot of time, actually, on the project, vetting all this information. It's an enormous part of what they do. Beria sets up several parallel labs so that they can check and check and check.

Finally, it's all in English, which is a problem as well, and needs to be translated. That's also a very difficult issue.

In addition, there is a parallel project within the Soviet Union, set up by Beria, of Germans who are recruited from mostly the eastern part of Germany in June and July of 1945. They are a backup plan. They are working on an atomic bomb as well, but they are doing it—they have no contact with the Soviets, with the one exception, and they are supposed to work the problem if the Soviets fail.

In the end, the Soviets make a device, which they call RDS-1. What "RDS" means has a lot of debate about it. But the only way they know that it works—the only way you can ever know that the nuclear device that you have works—is by destroying it. You have to explode it in a nuclear test. Otherwise, you don't know.

So that's what the Soviets do. They take their one bomb that they have and, on August 29, 1949, they detonate it at a test site. They call that episode "First Lightning". It's one of the most extensively engineered test sites for a nuclear blast of that time period. They build subway tunnels underneath the ground of the test site at different distances and different depths to see how far a blast would go. They chain goats and other animals to tanks at different distances, planes at different distances, build bridges, to try to see what the blast effects of it would be, which is something the Americans don't start doing, really, until the 1950s.

Narrating the moment of that test and what the Soviets thought about it is a key part of the book, as is how the Americans find out about this test, which they do by filtering dust that's wafting off the Siberian coastline into the North Pacific, with a very strange system they set up to try to detect this blast. They end up being successful, and in the end, it's Harry Truman, on September 23, 1949, who announces the Soviet bomb, not Stalin.

The next issue, especially relevant for here, is the ethical implications of nuclear weapons work. What did it mean to make a bomb for Stalin? I don't mean geopolitically; I mean ethically. There is an awful lot of ink that has been spilled on the ethical implications of trying to make a bomb for Hitler. There's a huge amount of stuff on this, and Heisenberg gets extensively castigated for being willing to cooperate. Oddly, there is almost no discussion of what the ethical implications are of giving a bomb to Stalin, who is a dictator of no less murderous intent or span. No one talks about it.

What's most peculiar about this is that these scientists actually succeeded. They actually gave the dictator a bomb.

So I devote considerable time in the book trying to sift through moral reflections of the scientists about what the exact nature of their work meant to them. It's actually very hard to figure this out, because the reflections aren't written down at the time. All of them are written after 1953, when Stalin dies, and many of them are written in memoirs or in letters that are only released or edited after 1991, when the Soviet Union collapses. But you can glean something from them.

What do you glean? Basically, there is no moral hand-wringing at all. The reason there is no moral hand-wringing is because the Americans have already solved this problem for them. The Americans made the bomb first. If there was a moral decision about whether you should create a bomb, according to the Soviet scientists, that moral decision has already been solved; the Americans have made it.

The issue is whether you use it, and they don't plan on doing that. In fact, all they know is that the Americans have used this bomb, and they have used it against somebody who doesn't have a nuclear weapon. Every Soviet physicist brings up the issue of nuclear deterrence and stability and mutually assured destruction. This is what Andrei Sakharov says. This is what all of them say, from very early, from the late 1940s. As far as they are concerned, building a bomb is a morally unproblematic issue.

What I find particularly interesting is that the Germans who worked on the project also leave their moral reflections. Most of the Germans left some aspect of their moral reflections. Two of them stand out, Manfred von Ardenne, who becomes a leading figure in East Germany, and Nikolaus Riehl, who is the child of a Russian mother and a German father, born in St. Petersburg around 1900. He is bilingual. He is the most highly placed German in the project.

Both of them leave moral reflections. What's interesting about them is that they both worked on the Nazi project, as well as on the Soviet project. Yet, for some reason, they are not castigated in the German literature at all, as if the Soviet project cleanses them of having worked on the Nazi project.

They both have extensive thoughts about culpability. As far as they are concerned, the Nazi project is actually the most moral one because it fails, and the American project is the most culpable one because it leads to death. The Soviet project is somewhere in the middle there. And they all cite the nuclear deterrence issue.

Part of the point of this discussion, to my mind, is how frightening to scientists around the world the American actions at the end of World War II look and what their stake is in this period. There is a tremendous amount of uncertainty and agony among Soviet scientists, who have lived through the destruction of most of their major cities in World War II, to imagine that happening again.

The period of the monopoly is also full of anxiety for Americans, because they have no idea how long this thing is going to last, and they are actually terrified of what the Soviets are going to do. It's a period with no stability for anyone.

So what's the impact of "Joe 1?" "Joe 1," as Joanne said, is the semi-humorous name given to "First Lightning" by the Americans in reference to Joe Stalin. It's often represented in histories of the Cold War as a radical break of sorts. There is the world before "Joe 1" and then there's "Joe 1" and everything is different. It both is a radical break and it isn't.

What I do in the book is show how "Joe 1" is not so much a game-changing total rearrangement of how the world works, but rather it provides a new potent argument for people who want to escalate the Cold War on the American side. That is, certain people, like the physicist Edward Teller and Atomic Energy Commissioner Lewis Strauss, use the fact of Soviet proliferation as an argument for pushing policies forward that they had been advocating for a long period of time, quite successfully. I'll give you three examples of that.

The first is the crash program to build a hydrogen bomb, which is in late January 1950. Truman announces that we are going to go all-out and attempt to make a thermonuclear weapon. Thermonuclear bombs have theoretically infinite scalability. They can destroy much larger areas of terrain, orders of magnitude higher than a fission bomb.

There is an extensive debate, starting from 1942—but it heats up very much around the time of "Joe 1," especially in October 1949—about whether the Americans should embark on this program. Scientific opinion is split, but mostly in favor of not doing it. But with "Joe 1," that changes. There is a very strong argument that the Soviets have made a nuclear step and we must respond with an additional nuclear step that is more extreme.

One can ask what would have happened had the Americans not built their hydrogen bomb. Almost certainly what would have happened is that the Soviets would have built one anyway. But it's clear that the hydrogen bomb is, in many ways, propelled forward, the escalation of the arms race is propelled forward, by "Joe 1," on both sides.

The second issue is anticommunist hysteria. Right after the Truman announcement on the H-bomb, on February 4, Klaus Fuchs is publicly arraigned for treason in England. The Venona codes—they are later called Venona—encrypted telegrams that are sent from the U.S. back to the Soviet Union, pick out a man called "Charles" or "Rest." That person is Klaus Fuchs. He is a very prominent theoretical physicist who was at Los Alamos, who passed some very important information along. He confesses in January 1950 and he is tried. That happens right on the heels of Alger Hiss' second trial, where he is convicted of perjury.

On February 14, 1950, in Wheeling, West Virginia, the junior senator from Wisconsin, Joe McCarthy, gives a speech about card-carrying communists in the State Department. "Joe 1" becomes part—the argument is, Fuchs gave them the bomb, hence we have "Joe 1," hence the world is much less stable, and we are infiltrated at every level. It becomes another piece of ammunition in this argument.

The third case—there are many more—is proxy war in Korea. This is on the Soviet side now. Kim Il Sung approached Stalin and asked whether he could now finally reunify Korea, now that the Soviets have an atomic bomb and there's no risk of nuclear retaliation. Stalin knows that he has no atomic bombs anymore, because he blew up the only one he had, but he authorizes it anyway. That starts the Korean War. There is a lot of evidence—although it's still, without the North Korean archives, unclear—that "Joe 1" is a major motivation for the North Korean attack.

So this post-"Joe 1" Cold War is the Cold War you remember. It's a Cold War of nuclear symmetry, a Cold War of an arms race, which was not the way the Cold War was structured before then, the one which is centered around Berlin and has 1948 and the Berlin blockade as kind of its peak.

In conclusion, what are the implications for today? What does this have to do with how we think about nuclear weapons now? I actually think quite a bit.

Although our mechanisms for detecting foreign nuclear blasts have changed radically over the period of time—and that's necessary because the system we used in 1949 to filter the atmosphere doesn't work when your nuclear bombs are detonated underground—the systems have advanced a lot. But all that does is add more data. It doesn't do anything to resolve the problems of the unreliability of information, the mirror-imaging, and the remote entity problems that are epistemologically grounded in the issue of foreign nuclear weapons programs.

If you take the quotations, for example, from the late 1940s, about how long until the Soviet Union gets a bomb and do a search-and-replace of the word "Russian" and put the word "Iran"—"Russia" and "Iran," just separate them—you get almost exactly the same articles. Part of the reason for that is that the question is not predicated, in philosophy-speak. That is, no one says what it means. How long until what? Two years until what, exactly? Two years until a closed fueled cycle? Two years until a test? Two years until a workable bomb? Two years until a deliverable bomb? Deliverable where? A stockpile?

These are all very different questions. When people throw numbers around, like 50 years or ten years or two years or six months, they never usually tell you what they mean. The implications of that are actually quite dramatic for policy. That's an error from the late 1940s that you see replicated again now.

I don't want to castigate intelligence people for doing this. Nuclear intelligence is extremely difficult work. There is just no way to get around these particular problems. They are really hardwired into the issue. The Americans still have to make policy decisions based on incomplete evidence, and so does everybody else.

Knowing the history of how we got to this place, I think, proves helpful, especially in honing the questions we ask about various different problems, such as Pakistan's nuclear arsenal, North Korean brinksmanship, or the status of the Iranian nuclear program. The questions are actually the same as what they were in the 1940s. Maybe this time the answers will be different.

Thank you for your attention. I look forward to your questions.

QUESTION: How do we know that Israel has a bomb? How do we know that South Africa has a bomb?

MICHAEL GORDIN: The issue of Israel is actually very interesting. It's the only case where we don't have a clear test. Except for the Israelis, everybody, who has a bomb has detonated it. That's how we know they actually had a bomb. The Israeli case is tricky for two reasons, and they both have to do with South Africa.

The South Africans actually tested a bomb. It's unclear whether they tested their bomb or they tested an Israeli bomb. But a bomb was tested, and it was a cooperative project in many ways.

There is another instance where there is a flash on a satellite. No one really looked very carefully at that moment. The South Africans were the people on the international control board who were looking at this site at the time. So it's unclear whether they have one.

The reason we think they have one is because of testimony from within the project, especially Mordechai Vanunu's revelations in the late 1990s about what happened at the Dimona reactor. From the 1960s onwards, from Kennedy onwards, there is tremendous anxiety about the Israeli civilian nuclear program and diversion from it. Enough testimony has leaked from within the program that we are reasonably certain—in fact, very certain”that the Israelis have a bomb. But that's a case where you get information from the inside that is mutually corroborated and clearly upsetting to the Israelis. So that gives you a sense that it might be true.

The South Africans tested something and then they dismantled it. It's one of the few cases of someone dismantling a nuclear arsenal when they had one. We know that because of the test. We also know that from the tremendous numbers of documents that have been released in the 1990s about what they did, and the process of transparency. But as far as I know—and I could be wrong—they no longer do have such a bomb. It was a fear of the apartheid state that there would be an African bomb. They didn't want that, and so it got rid of the program when it became clear that they were going to have to cede power.

QUESTION: I'm sure you are going to get lots of questions on the contemporary implications of proliferation, so I'm going to go back to history, if I could, for just a second.

MICHAEL GORDIN: Nothing would please me more.

QUESTIONER: On December 8, 1953, less than a year after he had been inaugurated, Dwight David Eisenhower gave the "Atoms for Peace" speech at the United Nations. Why?

MICHAEL GORDIN: This is quite a good question. There are a couple of good reasons why "Atoms for Peace" becomes a hot issue in the mid-1950s. The major reason is that there is an implied promise in the establishment of the Atomic Energy Commission in 1947 that there is going to be something to cleanse the bomb. The bomb is not just going to be a feature of war-making; it's also going to give something back. It's going to make energy that's "too cheap to meter," in the parlance of the day, and they are going to spread it around the world in peaceful civilian form so that no one will want to make a program of their own.

There are two reasons why this is done. The first is that it's an anti-proliferation measure. It ends up being counterproductive for that purpose. But the idea was, if we give people cheap energy and give them civilian nuclear reactors, they won't develop the expertise to develop their own bombs and they also won't weaponize in secret. We will have inspectors and people on the ground.

The major reason he does it is because he wants to ramp up the American nuclear arsenal dramatically. It's incredibly expensive. Eisenhower is quite worried about the problem of destroying the village in order to save it. He is worried that the arms race is going to bankrupt the United States before it bankrupts the Soviet Union. He doesn't want to, therefore, build up a huge conventional force. He wants instead to transfer a huge amount of our deterrence capacity to nuclear weapons. They are much cheaper to build, especially at that period of time, especially with thermonuclear weapons. In order to do that, he is using "Atoms for Peace," in many ways, as a public relations ploy to mask what the new look, the policy of aggressive deterrence with nuclear weapons, is doing.

It takes the wind out of the sails of the Soviets, actually. They are quite upset about this. They issue their own initiative of "Atoms for Peace" roughly around the same time, in clear, direct response to Eisenhower.

Those are my readings of why "Atoms for Peace" is initiated.

Do I also think that Eisenhower believed that atomic energy was going to be revolutionary and transform the Third World into modern nations? Yes, he believed both aspects of that. But the timing of it and the way it's delivered have the elements of political ploy. There's no reason why things can't be both.

QUESTION: How do you control and stop the spread of nuclear weaponry, particularly to terrorists? How would you deal with the problems of Korea and Iran and Pakistan?

MICHAEL GORDIN: Those are extremely large questions. I should say again that I'm a historian. I can give you informed speculation on these. I can especially give you informed speculation on the control part.

Most efforts of controlling nuclear weapons, up through, I would say, the 1970s, focus on controlling information, trying to keep the secret balled up. That actually doesn't work. Way too much information is open, accessible, and can be put together by an informed observer, such that even the Soviets using publicly available information get a very long way towards making a bomb and would have probably managed, in not much longer, to make the bomb without spy information. There are also lots of spies running around stealing information. Information control doesn't work.

There is one thing that you need to build both fission bombs—that is, the first kind of bombs—and thermonuclear bombs, and that's access to uranium. That's the only bottleneck in the process. If you really want to control nuclear weapons, you can't try to do so by bottling up secrets and you can't do so by sequestering scientists who know something about it, because that knowledge can get out. What you can do is control the uranium—follow it around and try to keep a lock on it.

There are lots of ways one could try to do this. At the moment, uranium mining is also a business, as well as being a tremendously dangerous international activity. You would have to regulate uranium mining incredibly intensely. This is the idea behind the Acheson-Lilienthal Report of 1946, which becomes the Baruch Plan in June of 1946, the first real arms-control policy. It doesn't go anywhere at the time, for a variety of historical reasons.

But my suggestion is, focus on the uranium. That's the material that needs to be controlled. If there were no uranium, or if the uranium were highly regulated and we now can determine where it is on the earth's surface; it's quite common—we can regulate where it is, and especially access to highly enriched, very rich ores, like those in the Belgian Congo. That would be the mechanism for controlling it.

As for what happens, to deal with things later, the most important way of keeping it out of the hands of terrorists, I would say, is to secure the former Soviet arsenal. That's the most dangerous concentration, especially in Siberia, of ready-formed nuclear weapons.

We don't have to worry very much about a terrorist group building a nuclear bomb of their own. It's incredibly expensive to build a nuclear bomb. It's very, very difficult. It takes a long period of time, and it's hard to do undetected.

It is, however, reasonably easy to steal one, if you know where to do it. That means securing those arsenals, such as the former Soviet arsenal, which has now all been repatriated to Russia, as far as we know— although it's not fully documented—and the Pakistani arsenal. Those would be the most important parts.

The questions of how to deal with North Korea and Iran are beyond my capacity to answer. I would feel very hesitant in this room. But it is, of course, the million-dollar question.

QUESTION: This might be anecdotal, but there was talk about Eisenhower planning to use an atomic bomb in North Korea. Is there anything in the records about that?

MICHAEL GORDIN: Yes. He did and Truman did. The atomic bomb was always on the table at each of these periods of time. It was decided in the end not to use it, for both instances. But the threat of doing so definitely has aspects in accelerating the Chinese program to develop their own nuclear device.

In the Quemoy-Matsu crisis of 1955, there is an additional nuclear threat. There is a lot of nuclear saber rattling that happens. The Americans, the British, and the French do it quite a lot. The Soviets do it very rarely. They did it with the Suez Crisis.

But everyone does nuclear saber rattling in those periods. Eisenhower certainly contemplated doing it. There are plans afoot to bomb Cuba during the Cuban Missile Crisis. It's a weapon that's actually in the arsenal. It's not taboo. There are constant war plans to deploy it in lots of different situations. It hasn't, obviously, been used—used in wrath, I should say. Tests are their own form of public posturing. But it hasn't been used against civilians since August 9, 1945. But it is definitely contemplated.

One of the issues that Truman, at least, considers is that he doesn't want to use it in Korea, because then it would always be used against Asian populations. The same is true in Vietnam. It has a very bad optical effect. It also would have been disastrous for other reasons.

But it is definitely contemplated. And at that that time, there was some worry about Soviet escalation—whether, if they did use one, the Soviets would, in fact, try to respond by defending the Soviets with their own arsenal.

QUESTION: We have been in an era now of building-down between the U.S. and Russia. Do you think that both countries will continue to have a mutually-assured-destruction policy for the foreseeable future? If so, what can you say about the limits down to which they can go, but not further?

MICHAEL GORDIN: Certainly we are at mutually assured destruction now and at the current rate, it will take a long time to get to a point where there isn't mutually assured destruction. The question basically breaks down to how much of a threat you see Russia being and how likely you think it is that Russians would respond to American actions with a nuclear strike of their own. I personally think that risk is very low. The Russians have many better weapons at hand, such as closing gas pipelines to Western Europe. They have a lot more ways of influencing policy than nuclear weapons.

I would hope the build-down would continue. I don't think it will ever go to zero. I think there are some countries that might give up their weapons. The Russians and the Americans will never do that, for a variety of historical and actual strategic reasons. But I think that one can go a very long way and still maintain a credible nuclear threat.

Mutually assured destruction usually requires your ability—given an all-out Soviet strike on the Americans, the Americans, even if many silos are destroyed, can still manage to destroy everything else on the other side. We are not going to be at that capacity, but we will be at a capacity to destroy Moscow, and that's quite enough for deterrence purposes, in my view of how things are. I don't think Putin is insane or Medvedev is insane, but I also don't think we are insane.

QUESTION: Immediately after the war ended in August 1945, there was, I believe, a cabinet-level debate on whether to share the science with the Soviets. Truman's advisers were split. He decided not to share the science. Do you think that that had any material impact on relations with the Soviets? Would they have been more cooperative if the science had been shared, or did they have the science anyway, on a general level, and it wouldn't have made a difference?

MICHAEL GORDIN: What you are referring to is the so-called Stimson Plan, which is first articulated right after August 15, 1945, while Stimson is on vacation. He is the secretary of war, and in World War II he is quite elderly and needs rest. But he is very worried about this. On his last day in office, which is, I believe, September 21, 1945, they have a cabinet-level meeting to discuss this. People are split three ways. There is a Forrestal faction—Forrestal is the secretary of the Navy; he becomes the first secretary of defense—that says absolutely not. There's another faction, which is the Stimson faction, which says we should give some information. But the usual way it's put is, we give them the basic science, but not the engineering issues to build a bomb. Then there's the Henry Wallace faction, which says we should give everything.

Truman decides not to do anything. He doesn't side with Forrestal exactly. He just says, "I'm not going to solve this problem yet. I'm going to discuss it with the British." And the Acheson-Lilienthal Report comes out of those discussions.

Would it have made a difference? In terms of the science, they had that already. What was being proposed to give was basically the kind of information that was in the Smyth Report or a little bit more. They would have already had that information or developed it very soon on their own.

That issue, though, doesn't bear on the heart of the question you are asking there, which is, would it have changed Soviet behavior? It's possible that some gesture of openness would have maintained the rocky but somewhat cordial relations that existed between FDR and Stalin. But, for a variety of reasons, I think that what happened at Potsdam and later had already damaged relations to a point that it wouldn't have been likely to make Stalin more conciliatory, even had Stalin been willing to be more conciliatory, which I frankly doubt, based on material from the archives.

In particular, there is this episode—it's how the book starts, actually—this famous episode where, at the conference table at the Cecilienhof Palace in Potsdam, during a break, Truman saunters over to the Soviets and says, "We have an explosive device of extraordinary magnitude."

Stalin says, "Good. You should probably use that."

He doesn't say the word "atomic." Of course, Stalin knows it's an atomic bomb. But what he knows now is that Truman doesn't want to use the word "atomic" around him, so already he doesn't trust Truman. So the very fact that something was revealed, but not enough was revealed, heightens suspicion and wariness on Stalin's part. That would probably have been the same case in September of 1945 as well. Some information would have been given, and Stalin would think, "This is a trick. They haven't given us the right stuff or they're giving it to us for some other reason that I can't figure out just yet." It probably wouldn't have helped.

QUESTION: For all the devices that exist, have any studies ever been made of how many of them could actually be set off without causing a cascading deterioration in the atmosphere itself?

MICHAEL GORDIN: The answer to the question is, yes, the studies have been done, but they are all simulations. The largest single bomb was detonated 50 years ago, plus a little bit, called the Tsar Bomb, which was a 100-megaton bomb that was only half packed. It was 50 megatons. It was detonated over Novaya Zemlya, which is the island off the coast. It didn't detonate the atmosphere. It's the biggest thing anybody has ever designed. No one designs anything that big. So that's the best we do.

We could probably detonate a lot of them, to have a bigger megatonnage than that. We can certainly beat 100 megatons or 50 megatons by detonating several. It could, in principle—there's a low probability—detonate the atmosphere. But we don't know the answer to that.

We could have a huge blast. The question is whether it would ignite the atmosphere or whether it would create a nuclear winter. Igniting the atmosphere has been simulated with computer models. We have gotten a lot better at simulating models of the atmosphere. The argument that one sees in the classified documents that have been declassified—and not—is that it wouldn't set off the atmosphere. But that's because it's a low-probability event. That doesn't mean it wouldn't happen.

The real question is, would any of our bombs still work? We haven't tested one since 1991 or 1992. The Russians haven't either. So we actually don't know, after about 20 years, whether the glue still holds, how much deterioration there has been. We just don't know. We do modeling, again, with computer simulations, detonating the explosive parts with the fissile core removed, but we don't actually know whether any of our arsenal would actually work if deployed. That's what most people are worried about at Los Alamos and Livermore, not the atmosphere detonation.

But there's no way of knowing unless we really try it and detonate the entire arsenal, and I think no one wants to do that.

QUESTION: In the 1950s, there was a lot of concern about a Soviet invasion of Western Europe, and American policy, as I understand it, was to be prepared to respond with first use of nuclear weapons pretty quickly if that happened. That was felt at the time, I think, as sort of the most likely place for a nuclear war to break out. In retrospect, like most deterrence—the problem with successful deterrence, of course, is that if it's successful, it's not clear if there is a threat. The argument has been, well, there never was a real threat.

Do you have any sense, from looking at the documents, of how serious that threat was and at what times, in particular, it was more serious than others?

MICHAEL GORDIN: My own personal view from the documents—part of the problem is that Stalin, like most people with a large army, war-games out all possible options. So there is obviously discussion about invading France. I don't think it was very likely, even at that moment. And that would have been the moment that was most tense. 1945 to 1948 would be the most likely period of time. But I don't think it was ever very likely. I think there was an intense desire to hold onto the territory they had—and possibly at very specific moments, like very late Khrushchev years, when he is quite tense and his power base domestically is unstable, of trying to get some of the rest of Germany back.

But that's my qualified guess. I don't think it's a likely option. I think, in many ways, containment worked, or at least the Soviet Union decided it was not worth trying very hard to expand beyond those levels. In 1955 is when they leave Austria. If they had wanted to, they could have stayed. There are lots of other instances where they leave Iran, they leave lots of places where they feel like it's not worth holding onto for the sake of maintaining what they currently have.

QUESTION: You mentioned earlier that there were all kinds of estimates about how long it takes. What's your estimate about how long it would take Iran to come up with a real bomb?

That's only part of the problem. The other part is, if in fact the craziness took hold and it was bombed, how long would it take to rebuild a bomb? In other words, does that make sense from the point of view—you bomb them and then they have the wherewithal to do it all over again.

Last of all—this is, I guess, a science question—when you blast a nuclear site, are you unleashing anything in the air?

MICHAEL GORDIN: To the first question, I can't give you an estimate. The reason why I can't give you an estimate is that I don't know Farsi. I don't have access to classified documents. I can just do what you can do, which is read the newspapers, government reports, and make some sort of assessment. It depends again on what you mean. How long until they have a fuel cycle and enough enriched uranium to make a device? That's probably not very long. How long until they weaponize that into a deliverable device? It's a lot longer. That's much harder than just having the fuel. Would they decide to test or not? That's an additional issue. How long until they built a stockpile is quite a while. So the question is what you think they might want to do with such a device. Do they just want to test it to say that they have it, which is what India did in 1974, or do they want to use it for some purpose?

So I'm going to respectfully duck the question of estimating how much time. I think I can say that it's a very difficult question to answer, for all the reasons that we don't have access to lots of secret sites, we don't know who we can trust, and we have lots of reliable and unreliable data that is coming in, and it's very hard to filter them out.

The second question is, how long does it take to rebuild? No country has ever abandoned their nuclear weapons program because of military force. The one case that is a possible qualification is Iraq after the bombing of the Osirak reactor by Israel. But that didn't stop their nuclear program. It just slowed it down. They had it rebuilt by 1991. That's not very long. It's seven or eight years to rebuild. So, in general, that's not true.

Nations have, in fact, given up nuclear weapons programs, all the time. Switzerland had one in 1946-47. They stopped. Sweden stopped. Libya stopped. South Africa gave theirs up. Ukraine and Kazakhstan gave theirs back to Russia. There are lots of people who have given it up. Brazil and Argentina both decide to back out of the race. The first nation that decides that they are absolutely not going to have a program was Canada, which was actually a partner in the Manhattan Project and could have done so very easily.

So lots of people have given up nuclear weapons over time. No one has done so by force. But one could say there's a first time for everything. That's just the historical fact of the matter.

If you bomb a nuclear reactor, of course you release stuff into the air, because there is lots of fissile material and it would spread. You would have something at least on the order of Three Mile Island, probably worse, if you did that, because the core would probably crack and there would be released plutonium spread around the environment. That's a danger. It's a different danger from a nuclear explosion, because the material wouldn't go critical. But it would spread and it would contaminate groundwater and it would do lots of bad stuff like that.

QUESTIONER: Could it be a Chernobyl?

MICHAEL GORDIN: It could be, depending on what you did. Probably not.

One thing you could do, if you try to bomb a reactor and you don't do it exactly right, is you can just destroy the cooling mechanisms so that the reactor goes critical. That would be a Chernobyl incident. So that could happen. But it would have environmental consequences no matter what you did.

QUESTION: Five, eight, ten years ago, there were many allegations of the Soviets having manufactured and then lost track of something like 52 suitcase bombs. I wondered if during your research you have come across anything that is authoritative about this, and where that might be today.

MICHAEL GORDIN: The short answer is, no, nothing authoritative. It comes up all the time, the allegation. But there is no reliable documentation either way on this. There are a couple of ways of explaining the lack of reliable documentation. Either they have actually lost track of it and they create lots of documents to indicate that they haven't in order to obfuscate the issue—because it's hugely embarrassing and dangerous—or, in fact, they actually haven't lost it, and it's just allegations.

I don't know the actual answer. I wish I did know. Securing the post-Soviet arsenal is, again, the single most important thing that would need to happen.

JOANNE MYERS: I want to thank you very much for a wonderful discussion.