Deep Black, by William E. Burrows
Ignorance vs. Intelligence
Deep Black: Space Espionage and National Security.
by William E. Burrows.
Random House. 401 pp. $19.95.
William Burrows, who teaches journalism at NYU, here performs a tour de force. He reveals the existence of nearly all U.S. technical-intelligence systems and describes many of their capabilities. Nevertheless, he manages wildly to misunderstand what these systems can and cannot do.
Although the book is about “black,” that is, highly secret matters, its shortcomings are not due to lack of data. Rather, Burrows does not understand his subject because he is the eager intellectual captive of sources in and around the U.S. intelligence community whose primary commitment in life is to advance the cause of arms control, and who therefore see their field exclusively in that distorting light. These sources fed him scraps of facts—many of them wrong—and much ideology.
The intelligence community has officially warned all those (myself among them) whose background would permit them to disentangle the real from the fanciful in this book not to do so, lest they compound the damage. Hence, I will largely refrain from quibbling with the book’s claims about the visual resolution or electronic sensitivity of individual satellites, etc., and concentrate instead on the perspective that leads Burrows to put his facts, both real and imagined, at the service of obvious political ends.
To begin with, the reader should understand that Deep Black is a journalistic rather than a systematic treatment of the subject. A systematic treatment might begin with a survey of the information that the U.S. requires about the Soviet Union for various purposes. Then it would describe the technology of intelligence, and show what that technology theoretically could and could not do to satisfy the various requirements. It might also describe how the U.S. technical-intelligence system has performed in the past. It might take note of how changes in the Soviet Union—its weaponry, its communications system, etc.—are affecting what we need to know and our ability to learn about it. Finally, it might examine the U.S. technical-intelligence system’s options for dealing with these changes.
But Deep Black is organized semi-chronologically. This approach encourages one to evaluate technology against itself rather than against the reason for which it was sought in the first place.
For example, how “good” was the U-2 aircraft that in 1957 brought back the first set of photographs of broad stretches of the central Soviet Union? The correct answer is the same as to the question, “How’s your wife?”: namely, “Compared to what?” The U-2 surely was a marvel of aeronautical engineering. It flew higher and longer than anything ever had done before. Its cameras were sharper and more ingenious. Above all, the U-2 greatly raised the level of facts at our disposal about the Soviet military. Similar things can and should be said about every one of our new aircraft-based and satellite-based sensors. Burrows’s stock-in-trade is to say them. But the effect is quite literally solipsistic—defining reality in terms of the ability of our technical systems to perceive it.
Burrows spends by far the largest part of the book conveying the impression that nothing of any significance, not even amorous affairs in the Soviet Union, can escape our technical wonders. Thus he states that the U.S. has not suffered a strategic surprise since Sputnik in 1958. Actually, however, Sputnik was no surprise, while in the years since, the U.S. has been truly surprised by the Soviet deployment of counterforce missiles under SALT I (not to mention the invasions of Czechoslovakia and Afghanistan, the ascension of the Ayatollah Khomeini, and the Yom Kippur War). It is most irresponsible, then, for Burrows to tell his readers that strategic surprise and surprise attack are now impossible.
To support this contention, Burrows cites the improbability that a complex of missile silos could be dug and loaded at any given location within the short period of time between “visits” to that location by a U.S. imaging satellite. He is deeply mistaken, because he fails to draw the proper conclusions from the facts that he himself points out: namely, that the U.S. has never had more than a handful of imaging satellites, and that the competition within the intelligence community for what they ought to look at is intense. From this alone it should be obvious that U.S. satellites mostly look only at a few of many places already known to be interesting, and almost never at anything else. Moreover, on those very rare occasions (so few that they can be counted on one’s fingers) when U.S. satellites have been able to see places not totally predicted by the Soviet Union’s satellite warning program, we have found the Soviets doing things with their missiles that are radically different from the things they do when they know they are being observed.
Indeed, after reading Deep Black, one would be surprised to learn that the U.S. does not even know whether the famous SS-18 silos are loaded or not. We have never even seen an SS-18 missile. Much less have we watched each and every silo as it was loaded, and then watched to make sure that the missiles were not subsequently unloaded and taken somewhere else. And Burrows also seems to have missed the fact that smaller missiles can be stored horizontally in warehouses, raised upright, and shot off quite without silos. Nor does he deal with the fact that the U.S. knows virtually nothing about how many or what kinds of missiles the Soviet Union is producing (since this activity goes on not in the open but indoors where it is invisible to satellites).
Even more important than imaging satellites are signals-intelligence (SIGINT) satellites. But they, too, “see” only the surface of things. For example: of, say, 100 channels of signals that we intercept from the testing of a Soviet missile, anywhere from 70 to 98 will often be in unbreakable code. Why did the Soviets choose the channels they chose for encoding? Why did they choose to broadcast certain others for us to read? Such questions are simply beyond the scope of our collection systems. With regard to modern phased-array radars whose signals we intercept and analyze, Burrows does not tell us that these signals are anything but the final word on the radars’ capabilities because both the signal that the Soviet radar transmitter sends out, and, more importantly, the use that the Soviet radar receiver makes of that returning signal, are constantly modified by the radar’s own computers. In other words, just as in the case of imagery intelligence, signals intelligence gives us only a few glimpses of Soviet reality.
Signals intelligence and imagery intelligence provide cues for each other. Burrows argues that what is not evident to the combination of U.S. imagery and U.S. SIGINT is not worth worrying about. But what does it mean to be evident to instruments that at best sample only the surface of a secret society? Here, Burrows’s answer is clear. “Over the years,” U.S. analysts have “categorized everything they’ve seen in order to establish precise operational patterns. What does not fit the pattern causes immediate suspicion.” Again and again, he emphasizes that the strength of our technical intelligence lies in this ability to recognize patterns. And why? Because the Soviets always behave according to rigid patterns.
The truth, however, is that the American intelligence system’s obvious reliance on “patterns” is not a strength but a fatal weakness that invites any sensible opponent to exploit it. Given that the Soviets have learned almost everything there is to know about U.S. technical intelligence, would it not be “cost effective” for them to manage their own exposure to it for the purpose of misleading us?
To take one example of many: in 1980 and 1981, the U.S. “observed” the Soviet Union using its newest surface-to-air missile system, the SA-12, to intercept its own test-fired missiles. Since that time, the usefulness of the SA-12 for antimissile work has been a staple of the domestic American debate over whether to build anti-missile defenses. Then in 1987 some CIA analysts told the New York Times that the SA-12 really should not be considered an anti-missile weapon because in recent years our electronic observations have revealed that every time the Soviets have used an SA-12 this way, they have missed. Yet when we consider that the Soviet Union’s paramount political-military objective in the 1980’s is to stop the U.S. from deciding to build anti-missile weapons, and that the Soviets know how we monitor their testing of the SA-12, is it not possible that they decided to use controlled miss-distances in their tests? Doing so would not really detract from the military usefulness of the tests and would yield a substantial political dividend, i.e., disinforming the U.S. about the Soviets’ own program of defense.
In addition to permitting Soviet deception, the reliance on patterns is conducive to self-deception. On the very rare occasions when U.S. intelligence analysts have been asked to investigate whether activities previously identified according to a pattern might conceivably be explained differently, they have found how thoroughly their colleagues had sometimes deceived themselves into not “seeing” very large and important activities that should have been obvious.
Few recognize that advances in technology are making this condition worse rather than better, and technical intelligence inherently less rather than more potent. Consider, for instance, how much simpler it is to gather intelligence about a missile than about a missile-killing laser. Missiles carry a load from point A to point B. If one can observe them doing this, one can learn quite a lot about them. But even the finest close-up photos of a laser device in operation will give no clues about vital parameters—such as pointing accuracy, jitter, evenness of the beam, fuel efficiency, retargeting ability—and few clues about the device’s power.
Only recently has the top echelon of U.S. intelligence admitted to itself that during the decades the U.S. was spending billions of dollars to monitor the width of silos within plus or minus 5 percent, the Soviet Union totally reversed the strategic balance and went from having 300 counterforce warheads to 6,000, largely within the letter of SALT I. How so? U.S. negotiators wanted an agreement to limit the number of Soviet missiles that could deliver counterforce warheads, but they realized that such a ban would be unverifiable, since American intelligence was unable to monitor the production of missiles or warheads, or the actual yield of warheads on any missile. Nor could our negotiators frame definitions of missile accuracy that could be monitored by U.S. intelligence. So they wrote the treaty in terms of features that U.S. intelligence could monitor—that is, launchers. But since the U.S. cannot monitor all ways of launching missiles, SALT I (and SALT II) only dealt with the two ways of launching which our imaging systems were capable of monitoring: silos and submarine tubes. More important, in order to be arguably verifiable, the treaties did not deal with actual missiles, warheads, or their potency. As we can now see from the proliferation of Soviet counterforce warheads, this “currency” that U.S. intelligence made available for the treaty turned out to be worthless.
Hence, since the mid-1970’s, there has been a frantic search in the arms-control community for a better currency. But this search is foredoomed. Yes, tomorrow’s cameras will be sharper than today’s, and some day our satellites may even carry most of the devices Burrows says they already carry. But no responsible person has suggested that we could seriously define the characteristics of anti-missile weapons in ways that would allow them to be monitored even by the sensors that Burrows describes.
Deep Black is going to be an influential book. Unfortunately, all too few readers will be able to verify just how deeply ignorant a book it is.