Code Warriors

by Stephen Budiansky

  If the Army wanted professional cryptologists it would have to create the profession itself and offer the kinds of rewards that customarily accompanied a professional career; it also would have to provide nearly all of the required professional training. In the war it had been possible to secure the temporary services of senior men, but a permanent organization would have to be built, and continually rebuilt, from the ground up with young men and women at the very outset of their careers. As Rowlett pointed out, the training demanded of a cryptanalyst “requires time and the resiliency of youth. The long hours of false leads and blind alleys demand the faith in ultimate success which characterizes young people.”36

  —

  The new and deliberately vague name for what had been the Army’s Signal Intelligence Service hinted at a more fundamental difficulty the wartime signals intelligence agencies were encountering in trying to remake themselves as a permanent part of the government national security establishment. (The Navy matched the Army Security Agency with its own minor masterpiece of bureaucratic obfuscation in the new name for Op-20-G’s Nebraska Avenue headquarters: Communications Supplementary Activities, Washington, or CSAW—pronounced “seesaw.” The British Government Code and Cypher School, which in April 1946 had moved from its never-convenient emergency wartime location at Bletchley Park to a set of drab government office buildings in Eastcote, a northwestern suburb of London, was rechristened with an equally misleading name, Government Communications Headquarters, or GCHQ.) The very existence of government bureaus that were intercepting and decoding communications in peacetime was a matter of moral murkiness, with the constant potential for diplomatic, legal, and political backlash. Instructions that ASA issued in 1946 decreed the use of the new code word CREAM for “special intelligence,” defined as all information “which results from the decryption of the texts or substance of encrypted communications.” The aim, the order explained, was “to restrict dissemination of this type of ‘TOP SECRET’ material within the narrowest possible limits, based on the ‘need to know.’ ”37 It pointedly added:

  Disclosure of “CREAM” material or its existence causes a twofold embarrassment to the U.S. Government:

  (1) Operational and loss of intelligence

  (2) Political

  The lack of any clear statutory authority for what they were doing reinforced the imperative to disguise their work. Wartime censorship had provided adequate legal authority for the military agencies to copy cable traffic, and a lengthy brief prepared by the Army’s judge advocate general at the end of the war argued that the president, by virtue of his broad powers to conduct foreign affairs, had the authority to continue monitoring foreign communications in peacetime; the brief also cited a 1928 Supreme Court precedent, Olmstead v. United States, which held that government wiretapping did not violate the Fourth Amendment’s privacy protections, since it did not involve the “searching” or “seizure” of a person or his private effects or the physical invasion of his home.38 Getting around the language of the Federal Communications Act of 1934, however, required more dexterous legal footwork, since the act pretty clearly made it a federal crime for anyone to divulge to a third party the contents of any intercepted communication.

  The cable companies nonetheless agreed to continue their wartime arrangement, allowing an Army officer in civilian clothes to pick up microfilm or paper tape copies of all international telegrams. “The Army came to me and asked for the company’s cooperation,” the president of RCA Global related years later to congressional investigators looking into the origins of the program, “and, by damn, that was enough for me.” But in fact the entire arrangement had plenty of people worried, and ASA went to great lengths to keep it secret. The presidents of RCA Global and the two other major cable firms, ITT World International and Western Union, sought repeated assurances from the government that the program was essential to national security and that their companies would not face prosecution for their actions. In 1947 the executives called on the new secretary of defense, James V. Forrestal, to ask if he could reassure them that the president himself had authorized the program. Forrestal did, though he added that he could not bind his successors.39

  Yet putting things on a more solid legal footing risked exposure itself, and when a plan was discussed by the signals intelligence agencies the following year to submit formal legislation that would legalize the companies’ assistance, the counsel for the U.S. government’s interagency Communications Intelligence Coordinating Committee warned that the crucial thing was to head off any “embarrassing line of questions” when the bill was taken up by Congress. “The greatest danger in debate will be ill-advised screams from some members of the press that the bill invades freedom of the press and civil rights.” The bill died a quiet death, and the question of the legality of signals interception in peacetime retreated back into its murky shadows. It would fester there for another three decades, until the Watergate scandals at last made it impossible to ignore the hidden toll that excessive secrecy exacted, which by then had become every bit as damaging as the much more obvious price to be paid by excessive disclosure.40

  *

  *1In English, for example, E has a frequency of about 12 percent, Z about 0.3 percent, versus the 3.8 percent probability that would be expected if all twenty-six letters appeared with equal frequency. The spelling system used in the Russian codes was actually more complicated, based on digraphs—a different number stood for each unique pair of letters, such as CK, EE, LM, and so forth—but the basic principle of cryptanalysis was the same, as digraphs have a characteristic frequency distribution just as do individual letters.

  *2Trying to solve messages with a depth of two where each message had been encoded with a different codebook, as was the case with depths between the ZET and ZDJ systems, was not completely impossible either, assuming enough progress in code recoveries in each codebook. But it was obviously taking things to a level of unprecedented difficulty. The Red Reverse discovery, though, meant that a sizable amount of key recovered from the reversed-page reuses within the voluminous ZET trade traffic might also be directly duplicated in some of the ZDJ diplomatic traffic, in effect giving a depth of three.

  3

  Learning to Lie

  Throughout 1947, Meredith Gardner continued to toil away on Arlington Hall’s most secret project, adding to the reconstructed codebook that lay at the heart of the Soviet spy messages. It was becoming clear that the four different one-time-pad code systems used for messages sent to and from Soviet embassies and consulates each belonged to a different organization of the Soviet foreign and intelligence bureaucracy: GRU, or military intelligence; naval GRU; NKGB; and “true” diplomatic, which dealt mainly with mundane consular matters such as visa applications. Up until 1941 the GRU was the main foreign intelligence organization of the Soviet Union, but the NKGB, on Stalin’s directive, had since eclipsed the military spies in that role, even though the GRU and naval GRU continued to operate abroad.

  Nearly all of the NKGB messages that had been rendered vulnerable to cryptanalysis owing to their encipherment with duplicated key pages were sent during just three years, 1943 to 1945. The Arlington Hall cryptanalysts would eventually conclude that the duplication was the result of wartime disruptions that forced Soviet printing plants to cut corners for a few months in early 1942. Once the duplicate pages issued during this short period were used up, the window on the Soviets’ most secret communications left ever so slightly ajar by this fatal slip slammed shut once again: even with the codebooks broken, only messages in depth could be read.1 But in the meantime thousands of potentially breakable messages enciphered with duplicate key had been transmitted—and successfully reading them might still yield valuable intelligence about ongoing Soviet espionage operations or the identity of agents still in place. When the project, best known by its final code name, Venona, finally ended in 1980, 49 percent of the 1944 NKGB New York–Moscow messages had been read at least in part, 15 percent of the 1943 messages,
and 1.8 percent of the 1942 messages, along with 1.5 percent of NKGB Washington–Moscow traffic and 50 percent of 1943 naval GRU Washington–Moscow traffic.2

  The Kod Pobeda book found by TICOM Team 3 was applicable only to NKGB traffic sent before November 1943, and as of 1947 Gardner remained unaware the book had even been recovered. Thus the effort at Arlington Hall at this point still relied upon pure cryptanalysis and “book breaking” to reconstruct both the one-time-pad pages and the underlying codebooks, all sight unseen of the originals. Gardner’s breaking of the spell table used in the post–November 1943 NKGB Jade codebook meant that messages containing proper names were the ripest for picking, and by the summer Gardner had recovered portions of several 1944 and 1945 messages that contained what were clearly Soviet cover names for dozens of agents in America. The results were so fragmentary that issuing them as individual serialized translations and circulating them to Arlington Hall’s usual list of recipients in the War and State departments and the White House, the practice throughout the war, seemed slightly ludicrous to Gardner, as well as all but impossible given the extraordinary security restrictions surrounding the work on the Russian problem. Not knowing quite what else to do, he decided to write up a series of special reports that attempted to pull together the available information; Gardner sent them to ASA’s Cryptanalytic Branch chief, Frank Rowlett, in the hope that he might get the attention of top officials at ASA and the War Department’s military intelligence staff, G-2, who in turn might know what to do with material that was at once too enigmatic and too hot to handle through normal channels.3

  A draft of Gardner’s “Special Analysis Report #1” landed on the desk of Colonel Harold G. Hayes, chief of ASA, on July 22, 1947. Gardner listed about two dozen cover names he had culled from hundreds of partially readable NKGB messages. From context he was able to determine that certain other words spelled out in those messages were cover names for places: TYRE was New York City; SIDON was London; CARTHAGE was Washington. In several messages the code word ENORMOZ appeared, once in juxtaposition with the names of Los Alamos scientists; it seemed likely that this was the code word for atomic bomb espionage. And an agent code-named LIBERAL and ANTENNA was mentioned as a go-between who passed on information gathered by certain other persons working on ENORMOZ. In one of his special reports Gardner summarized the information he had been able to assemble about this agent:

  LIB?? or possibly LIBERAL: was ANTENKO until Sept 1944. Occurs 6 times, 22 October–20 December 1944. Message of 27 November speaks of his wife ETHEL, 29 years old married 5 years.4

  Another agent who caught Gardner’s eye was G or GOMER or HOMER; he had cropped up in a number of Washington–Moscow and New York–Moscow messages. HOMER appeared to be well connected to highly placed British officials, and from the fragmentary translations Gardner was able to make in 1947 and 1948 he seemed to be supplying political intelligence of a distinctly above-average quality: in one message, HOMER provided details on advance planning for Roosevelt and Churchill’s secret conference that took place in Quebec in September 1944, while others proved to be verbatim copies of telegrams sent by Churchill to the British embassy in Washington.5

  Still other messages read by Arlington Hall in 1947 and 1948 unmistakably pointed to well-placed Soviet agents in the U.S. War Department and in the Australian government. Two messages from New York to Moscow sent in late 1944 on the NKGB system contained long quotations of documents supplied by an agent ROBERT: the documents, prepared by the War Department general staff and classified Top Secret, dealt with plans for postwar U.S. troop deployments. In October 1947, Gardner read two long cables in the same system that had been sent from Canberra to Moscow in early 1946; the Australian traffic had been intercepted by the British, who turned it over to Arlington Hall. They appeared to be extracts of a classified report published by the British War Office. GCHQ’s liaison man on the project, Philip Howse, contacted London to ask that a search be made for the original publication for comparison with the ciphered cable. GCHQ replied two weeks later: “Papers have been traced and being forwarded forthwith. Wonderful job.” This was a huge windfall, providing a spectacular run of matching plaintext—a “crib” in cryptanalysts’ lingo—which added a flood of codebook recoveries.6

  That the Soviets had spies operating in the Manhattan Project and in the highest levels of the British, American, and Australian governments was startling enough, but the traffic revealed something else about America’s erstwhile ally and rapidly emerging foe. The Russians were, to put it simply, extremely old hands at this game. To the Soviet spymasters conspiratorial thinking, counterespionage, and countersurveillance were second nature, and the precision with which they spelled out arrangements for dead drops, communications procedures, and shaking tails showed how thoroughly they grasped that the spy game was above all a deeper counterspy, or counter-counterspy, game. A message to a Soviet NKGB officer in London shortly after the end of the war specified the elaborate precautions to be taken for a rendezvous with an American agent, a member of the Communist Party USA who upon returning to Washington from his Army service had “agreed to continue work on the collection of intelligence”:

  DAN will go to the meeting and await our man for 10–15 minutes on the pavement immediately at the exit of the Regent Park tube station in Regent Street. DAN will have the magazine “John Bull” in his hand. Our man is to approach first and, after greeting him, will say, “Didn’t I meet you at Vick’s Restaurant at Connecticut Avenue.” To which this DAN is to reply—“Yes, Vick himself introduced you.” After this our man is to show DAN a small (19 groups unrecoverable) is to show an exact copy of this label. Then the two men will [talk] business. We recommend ALAN to contact DAN.7

  To a later generation brought up on spy novels all of this might seem old hat, but in fact these were the hallmarks of a professionalism of a kind that their Western adversaries were neophytes at by comparison circa 1948. The Soviets had none of the Boy Scout bungling and amateurish credulity of their sources that had so tarred the OSS. When Igor Gouzenko defected in Canada, Moscow immediately sent instructions to protect its most important agents in the United States from exposure, even if it meant breaking off most or all contact with them for months, even years, in order to preserve their future usefulness:

  Surveillance has been increased. Safeguard from failure: HOMER, RUBLE, RAID, MOLE, ZHORA, and IZRA. Reduce meetings with them to once or twice a month. Minor agents should be deactivated. Carefully check out surveillance when going to meetings, and if anything seems suspicious, do not go through with them.8

  The NKGB was playing a long game, for keeps.

  —

  As careful students of Russian history like the State Department’s George Kennan observed, the Soviet genius for conspiracy, paranoia, and secrecy had deep roots. Russian hostility toward and suspicion of the outside world was not merely a reflection of the Communist system but the stubbornly entrenched legacy of a village peasant society whose backwardness had endured for centuries after its equivalents in Western Europe had been swept away by the modernizing influence of Christian morals, cash-based markets, public education, the rule of law, a politically engaged landowning class, and growing national and linguistic identity. By the end of the nineteenth century even the most remote rural farm worker in Norway or England or France was tied to something in the world beyond family and clan loyalties. The Russian muzhik, up to and even after the time of the revolution, lived in a world that was not even medieval but prehistoric in its horizons. There was virtually no cash economy; most landlords were absentees utterly uninterested in their communities; most of the Orthodox Church’s priests were uneducated and corrupt; public schooling was all but nonexistent; peasant families farmed land in a primitive system of shifting allotted strips to which no one ever held permanent title, and thus had no incentive ever to improve. Law was limited to what the government forced one to do, mainly pay taxes and be conscripted into the army. There was scarcely even a sense of Russian
national identity or patriotism: the czars feared anything that might imply an allegiance to institutions other than the person of the czar himself, and so actively discouraged the most basic attributes of the modern state.9

  The result, as Kennan explained in what would become his famous analysis of Soviet intentions and thinking—the “long telegram” that he cabled back to Washington from the U.S. embassy in Moscow following Stalin’s belligerent Bolshoi Theater speech in February 1946—was that Russia’s rulers had always needed an external threat to bolster a government that had always been “archaic in form, fragile and artificial in its psychological foundation, unable to stand comparison or contact with the political systems of western countries.” During an earlier tour in Moscow, in 1936, Kennan had made the point by composing a report consisting entirely of sentences taken from the dispatches of the American minister in Saint Petersburg in 1853, during the time of Czar Nicholas I: “Secrecy and mystery characterize everything”; “nothing is made public that is worth knowing”; the Russian government possessed “in an exquisite degree the art of worrying a foreign representative without giving him even the consolation of an insult.” Or, as Kennan put it in his own words on another occasion, “Russians are a nation of stage managers: and the deepest of their convictions is that things are not what they are, but only what they seem.”10

  Marxist-Leninist ideology had only applied a crueler, pseudo-intellectual gloss to the venerable instincts of Russian rulers. To the “enemies of Russia” the Bolsheviks now added the “enemies of the Socialist revolution”; that ever-present threat, Kennan explained, provided them an ever-present justification “for their instinctive fear of the outside world, for the dictatorship without which they did not know how to rule, for cruelties they did not dare not to inflict, for sacrifices they felt bound to demand.” Concessions by the West would not alter Stalin’s behavior, because the Soviet Union had never been a normal state, interested in fostering international stability.11