The Emperor's Codes Read online

  When the Japanese Emperor Yoshihito died at the end of 1926, the official report of his death and the succession speech of his son Hirohito were relayed to every Japanese diplomatic, naval and military outpost around the world. Nave knew that the Japanese love of ceremony and obsession with predictable courtesies would ensure that every message was exactly the same. It was a simple task to follow it through the Japanese codes, breaking each in turn.

  The codebreakers in London concentrated on the standard cryptographic technique of examining the statistical peculiarity of the Japanese language to work out the probability of each individual character appearing. The first step in breaking any cipher is to try to find features which correspond to the original plain text. Whereas codes substitute groups of letters or figures for words, phrases or even complete concepts, ciphers replace every individual letter of every word. They therefore tend to reflect the characteristics of the language in which the original text was written. This provides the codebreaker with a relatively easy entry point. For example, the most common letters in the English language are E, T, A, O and N. If a reasonable amount, or ‘depth’, of English text enciphered in the same simple cipher were studied for ‘letter frequency’, the letter that came up most often would represent E. The second most common letter would be T and so on. Writing out the letters recovered in this fashion will reveal obvious words with letters missing, allowing the codebreaker to fill in the gaps and recover those letters as well. They can in turn be filled in, producing other obvious words and so the process goes on until the whole message can be read.

  Another basic weapon used by the codebreaker, ‘contact analysis’, takes this principle a step further. Some letters will appear frequently alongside each other. The most obvious example in the English language is TH, as in ‘the’ or ‘that’. So, by combining these two weapons, the codebreaker could make a reasonable guess that, where a single letter appeared repeatedly after the T which he had already recovered from letter frequency, the unknown letter was probably H, particularly if the next letter had already been recovered as E. In that case, he might conclude that the letter after the E was probably the start of a new word and so the process of building up the message would go on.

  The Japanese diplomatic messages were sent using romaji, where the syllables were spelled out using the letters of the Roman alphabet, and were particularly susceptible to contact analysis, one codebreaker recalled.

  The orthographic structure of the Romanized Japanese used by the Japanese Foreign Office in its telegraphic communications worked to our advantage. In this form of Japanese certain limitations applied to the letter Y. It was [nearly] always followed by A, and almost never by E or I. Pairs of vowels frequently occurred, and the most common were the combinations OO, UU, AI and EI in about that order of frequency. YUU and YOO often occurred, preceded by a consonant, in combinations such as RYOO, RYUU, KYOO and KYUU. It was our hope that we could use these characteristics of the Japanese language to make accurate assumptions for plain text.

  Nave made swift progress with the Japanese naval codes. ‘Using our own reporting messages as background information, I was able to decode all Japanese naval reporting code traffic towards the end of my first year,’ he said. He also made considerable inroads into the more complicated General Operational Code used for most major messages, which was sent in groups of nine letters.

  He soon built up a complete picture of the Imperial Japanese Navy, with every ship and its base recorded and every call sign and frequency they used for their radio messages carefully logged. As a result, the British were able to keep track of Japanese naval movements and managed to read a considerable proportion of the messages passing between the Japanese ships and their bases.

  At the end of each month, Nave sent a summary of the month's work back to London, adding the newly broken parts to each code, new additions to the Japanese order of battle and translations of key messages, as well as any material he had been unable to decipher, still on the red forms on which the Royal Navy wireless operators took them down. ‘The daily batch of messages kept me constantly busy – it was seven days a week – and I was adding to the code groups in my reporting code as well as constant additions to the list of call signs. I had complete understanding of their abbreviated addresses and, with additions to the call-sign list and with inter-flotilla and intra-flotilla information, I had a regular supply of information for my reports to the Admiralty.’

  The success that Nave and Hobart-Hampden enjoyed in breaking the Japanese codes and ciphers helped to encourage more interest in Japanese radio traffic. The Royal Australian Navy also intercepted Japanese messages and passed them on to Nave, and, in London, Harold Parlett, who had retired from his post as Japanese Counsellor in the British Embassy in Tokyo, was added to the GC&CS diplomatic section. Meanwhile, the British military based in Shanghai, the so-called Shaforce, set up its own interception facility as part of the Shanghai Army Intelligence Office, which operated from the British Consulate-General in the city.

  By the end of 1927 Clarke had decided that Nave's expertise was now such that he should be recalled to London to concentrate on breaking the Japanese Navy's General Operational Code.

  3

  A SPY BASE IN THE FAR EAST

  Nave arrived in London early in 1928, still totally unaware of the existence of Denniston's codebreaking operation. The Royal Australian Navy officer reported to his only contact in London, Rear-Admiral Sir Barry Domville, the Royal Navy's Director of Naval Intelligence. ‘He explained that I would be working outside Admiralty in an organization which operated under the pseudonym of the Government Code and Cypher School,’ Nave recalled. ‘At this time the entire staff of GC&CS totalled not much more than forty and they handled the breaking and translation of telegrams from all the leading countries.’

  Domville walked with Nave across Horse Guards Parade to the codebreakers’ new office in Broadway Buildings, Victoria, which they shared with MI6. They were taken up in the lift to the fourth floor where they were met by Denniston who told Nave that, partly as a result of his work, Whitehall was ‘beginning to take the Far East situation more seriously’.

  The Washington Conference had provoked a great deal of anger among the Japanese over their treatment and the apparent determination of Britain and America to keep them from achieving what they saw as their rightful place at the head of the world table. There was particular resentment at the British decision, under pressure from the United States, to terminate its treaty of alliance with Japan in favour of the far less comprehensive Four Power Treaty between Britain, America, France and Japan.

  Throughout the 1920s, the diplomatic telegrams deciphered by Hobart-Hampden and Parlett had shown a marked increase in Japanese militarism and a determination within the Imperial Japanese Army to expand its country's interests in China, where Japanese troops had been garrisoned as ‘advisers’ since the First World War. Japan was also pursuing aggressive trade policies, ‘dumping’ its own textiles on to the market in the British Empire, in direct competition to the Lancashire cotton industry.

  Whitehall was split over how to react to Japan's determination to flex its muscles as the leading power in the region. There was a considerable lobby for a more conciliatory approach towards Tokyo, but concern that if Britain backed Japan it risked antagonizing China and losing a substantial export market. Any intelligence that Nave and Harry Shaw could provide would be of immense assistance. The Japanese naval section was housed in a small shabby office containing an old desk, two chairs and a ‘secure’ cupboard. Within a few months, Shaw went out to the China Station and Nave was left on his own.

  My main task was to decipher and translate the telegrams to and from Admiralty Tokyo to Japanese naval attachés in Europe. These cables passed through British hands somewhere and I had copies of them on my desk every day. The daily pile of telegrams ensured I would have no spare time. The naval attachés had their own code and a codebook required building before this became readable. I
also had a steady research task as we had amassed quite a volume of messages in the General Operational Code. It was fortunate for me that I had available in the next room two eminent Japanese scholars, Hobart-Hampden and Sir Harold Parlett, who were always willing to come to my assistance when needed.

  By the end of 1928 Clarke was able to report that the Japanese Navy's General Operational Codebook was ‘quite legible’, thanks to Nave. ‘Over 800 messages in this book have been decoded during the year and the degree of legibility attained can be gauged from the specimen messages circulated to the Admiralty,’ Clarke wrote.

  The Admiralty had the Australian transferred into the Royal Navy and in July 1931 he and Shaw were rotated, with Nave going out to the China Station while Shaw returned to London. At the end of that year the Japanese Navy introduced a new 4-kana General Operational Code. It had a totally random order and there were two codebooks in use, one to encode the messages and another to decode them. But the codebreakers soon realized that, despite its apparent differences from the 9-kana system, it used the same basic codebook and they began to break it with relative ease.

  Two months after Nave arrived back in the Far East, Japanese troops based at Mukden in northern China staged a bomb attack, supposedly by Chinese dissidents, as an excuse to take over the province and create the puppet state of Manchukuo. The so-called Manchuria Incident provoked outrage not just around the world but in Tokyo itself, where the civilian government had been totally unaware of the army's plans.

  It also provided the first ‘live’ target for the British codebreakers since the end of the First World War. The Shanghai Army Intelligence Office and the China Fleet flagship party sent many bags of the ‘red forms’ on which the intercepts were taken down to London, where Clarke's naval section had been joined by a small military section. With Japan the only major military power actively threatening British imperial power, the intelligence the codebreakers produced was examined with increasing anxiety within Whitehall.

  The international condemnation of the Manchuria Incident led Japan to leave the League of Nations, further isolating it. By the end of 1934 it had occupied still more of northern China and abrogated the arms limitation treaties it had signed at the Washington Conference. Its foreign policy was now totally dominated by the Imperial Army, which held a virtual stranglehold over successive civilian governments and was determined to increase Japan's empire in the Far East regardless of what damage this might do to relations with the West. The extent of the danger posed to their interests in the Far East was brought home to the British in July 1934 when the codebreakers uncovered a major Japanese spy network centred on Singapore.

  Eric Nave had now returned to London and Harry Shaw was working in Shanghai as the head of intelligence and chief code-breaker for the Royal Navy's China Squadron. He deciphered a telegram from the Japanese Consul-General in Singapore to his bosses in Tokyo which revealed that the Japanese had two ‘top agents’ at the heart of the British Government in Singapore. They were apparently providing their masters in Tokyo with details of all the secret plans to build up the colony as a bulwark against Japanese expansionism, a ‘Gibraltar of the East’.

  Although enciphered inside the Japanese Consulate-General, the telegram had been transmitted by the British-owned Eastern Telegraph Company. Like all the cable companies operating on British territory, the company was obliged under the Official Secrets Act to supply drop copies to the authorities on request. By coincidence, it had been asked only three weeks earlier by naval intelligence to pass on any messages from the Japanese Consul-General, whom the British believed to be in charge of Japanese intelligence-gathering for the whole of the Malay Peninsula. It was immediately clear to Shaw that the request had paid dividends. The Consul-General's secret telegram to the Japanese Foreign Office, the gaimushoo, described what he claimed was a brilliant piece of espionage by ‘a very secret intelligence agent’.

  The deciphered message contained the main points made by Admiral Sir Frederic Dreyer, the commander of the China Fleet, to a conference on the construction of the Singapore naval base which had been held on board HMS Kent earlier that year. They included details of a secret submarine base, RAF trials for landing aircraft on the base's slipways, a military fort that was to be constructed on the island and the decision to set up a new, independent Royal Navy squadron based in Singapore.

  ‘England was extremely perturbed by the attitude of the Imperial Japanese Army and Navy towards the Far East,’ the Japanese spy reported Dreyer telling the conference. There was ‘absolutely no hope’ of the Japanese agreeing to a reduction of their naval strength. Given the strength of the Japanese forces in the region, and the state of the Royal Navy in the Far Eastern waters, the British could not afford to neglect their defences for a day longer. Work on the base was to be speeded up so that it would be completed within three years.

  At the end of his report the Consul-General explained how this explosive information had come into his hands.

  This was due to a secret person who had previously acted for me and who for a suitable reward obtained the minutes from a friend, who is a shorthand typist for personal confidential documents of civil Governor-general. This person took down the shorthand minutes of the conference. Hitherto, the shorthand notes this person has brought have been on the whole authentic. In this case, in addition to the bestowal of a large reward, I interrogated her. But she maintains there was no mistake. Actual date and names of those attending the conference will be obtained within a few days. The intelligence agent mentioned is said to be specialist in cartography and to be serving in public works department. He is now procuring a plan of the completed base very secretly. As I believe him to be amply reliable and it is necessary to pay him the reward that was agreed to, please telegraph me 1,000 dollars as special secret service fund.

  Shaw took the deciphered telegram to Dreyer who immediately forwarded it on to London, to Admiral G. C. Dickens, the Director of Naval Intelligence. Dreyer expressed his concern, but, while he curiously made no comment on the contents of the Japanese report, he doubted that it could have come from the naval conference itself since only commissioned officers were present.

  A preliminary inquiry decided that either the agents or the Consul-General must be inventing their intelligence reports for financial gain. MI5 was called in to investigate the affair ‘to be on the safe side’. But there was general agreement that British officers were scarcely likely to have allowed their female typists and secretaries to pass secrets to the Japanese.

  Dickens was not so sure. There was no doubt in his mind that Japanese espionage operations had increased dramatically over the past year. He had been complaining for some time at the decision to allow the Japanese to set up a ‘Naval Centre’ just a few yards away from the Broadway headquarters of Britain's own Secret Intelligence Service. The twenty-one officers based in the centre were known to be gathering intelligence ‘presided over by the Japanese Naval Attaché and thus to all intents and purposes sheltered by diplomatic privilege’.

  The increase in Japanese espionage operations was confirmed by the naval attaché messages being decoded by Nave. They revealed that a Royal Air Force officer, Squadron-Leader Major Frederick Rutland, had been recruited by the Japanese. ‘This sorry story commenced with a meeting, using assumed names, at a hotel in Manchuria,’ Nave said. Rutland was told to make contact with the Japanese Naval Attaché in London on his return to Britain. ‘We were unable to listen in to the plans made with the Attaché in London at their meeting in Richmond Park. However, the report made to Tokyo gave us the complete picture.’

  Rutland was to go to California to spy on the Americans. Nave passed the deciphered report to Sinclair and was called to a meeting with the head of MI6 and his deputy, Colonel Stewart Menzies. ‘The colonel asked if we were now to hand the matter over to MI5, to which he received the answer, “No, they will only make a mess of it, we'll handle this ourselves.” ’ Sinclair said the only evidence was the decoded messa
ges and these could not be made public for fear the Japanese might change their code. An MI6 officer was detailed to tail Rutland across the Atlantic where he was handed over to the FBI.

  The Japanese had also infiltrated a large number of spies into Malaya, bribing locally employed staff and even British servicemen to provide details of the naval base and the RAF facilities in Singapore. Both Dickens and Sinclair were concerned that the British, who still believed that there was something not quite gentlemanly about spying, were losing out in the intelligence war.

  The British attempts to break the Japanese codes were sufficient for peacetime operations but it was increasingly looking as if they might need to be ready for war. ‘The situation in the Far East has completely changed and left our intelligence arrangements high and dry,’ Dickens said. ‘The more I see of the cryptographic side of naval intelligence the more I recognize its vital importance and, further, that the nucleus we have at present is far too weak on which to build up rapidly an organization such as we would require in a war with Japan.’ He had begun recruiting new potential Japanese linguists and had only recently sent his deputy, Captain Campbell Tait, to the Far East to investigate the situation on the spot. It was Tait who had arranged with the Eastern Telegraph Company for copies of all cables from the Japanese Consul-General to be passed to Shanghai for Shaw's attention.

  ‘Shaw was specially sent to his present appointment because he is a trained cryptographer and a Japanese interpreter and therefore might be useful for occasional Japanese cryptographic work,’ Dickens said. ‘If it had not been for this and Captain Tait's initiative we should never have heard anything of this occurrence.’