Oversensitive Censorship

Just as it’s never difficult to overclassify a piece of information, it’s never difficult to find reasons not to declassify. The same principle governed the way official censorship operated during the Second World War: everybody knew that ‘Careless Talk Costs Lives’, and censors, responsible for second guessing what the enemy just might infer from an odd sentence, could be just as cautious. What follows is a diary entry dated 19 April 1945.

‘Are you security-minded? I mean, really security-minded? Yes; I agree the war in this theatre of operations is nearly at an end, but we must be careful. Are you as careful as this, for example?

The Films Division of the Ministry of Information sponsoring a brief film of the foolishness of unnecessary train journeys decided that the dialogue of a garrulous woman traveller relating her unnecessary journey to her sister, talking of Crewe station thus: "I never saw so many people in my life, if there was one Canadian soldier, there were fifty thousand," should be amended to, "if there was one Canadian soldier, there were thousands."

The official explanation: "Security reasons; just think how awful if there were fifty thousand."

Terrible; but perhaps they meant the crush on the platforms!’

(John Paddy Carstairs Kaleidoscope and a Jaundiced Eye Hurst & Blackett London 1946)

When Opposite Sides Share the same Crypto

I imagine that very few people think about the cryptographic consequences of a country breaking up but they can be significant, especially if the break-up is violent. The outbreak of the Spanish Civil War, for example, produced within days two opposing armed forces which used the same encryption systems which had to be modified on the fly by people with little cryptographic experience. The same thing happened in France in 1940 when the Free French military broke away from those who accepted the Armistice. One such break-up during the Second World War, that of the division of Italy into areas controlled by Mussolini’s Italian Social Republic based in Salò and those controlled by the Kingdom of Italy and based in Salerno, were written up after the war by Fred Catty, Head of the Italian Diplomatic Section in Berkeley St. (Commercial and Diplomatic Sigint were hived off from Bletchley Park to London early in 1942.)

Italian Diplomatic ciphers had been read without difficulty by GC&CS until in 1941 the Germans committed to joint operations with the Italians in North Africa. The Germans had been reading Italian systems themselves and simply refused to share classified material with their ally. However, the Germans decided that ensuring that the Italians gave nothing away of German intentions was more important than reading their traffic. They informed the Italians about the weakness of their systems and in July 1942 the old systems were gradually replaced by new ones. Fortunately for Berkeley Street the Germans did not bother to supervise Italian use of the new systems so they were never as secure as they should have been. In fact, by the end of 1942 the Italian Diplomatic Section was as successful as it had been before the new systems were introduced. In January 1943 the Italians appear to have had some misgivings and they added an extra level of security which should have brought GC&CS to a full-stop: but operator error neutralised the precautions the Italians had taken.  

In July 1943 Mussolini was deposed but escaped from captivity and set up his own Neo-Fascist Government in North Italy. In September the Royal Italian Government accepted unconditional surrender. Mussolini had his own Foreign Ministry located at Salò (Lake Garda) with its own WT Station while the Royal Italian Foreign Ministry was based at Salerno.

What follows is Mr Catty’s description (from HW 43/4 which hasn’t been released but which is quoted in HW 43/75-77, de Grey’s draft history, which has) of what happened next. It is worth highlighting that that the UK and US were as lax as the Germans had been in supervising their new allies’ security, and that the organisational and geographical separation of communications intelligence and communications security in the UK wasted cryptanalytic effort at Berkeley St.

‘There was quite an extensive array of posts which would claim to represent this Neo-Fascist Government abroad. In all of Europe occupied by the Germans the former Italian Ministers perforce remained Fascist; in addition, in Madrid, besides the 'Royal" Italian Embassy, Mussolini 'accredited' a Dr. Morreale (believed to have been the brother of Clara Petacci) as his 'Ambassador'. In Tokyo the Ambassador preferred interment as a 'Royalist', but Col. Principini, Military Attaché, took over. All these and isolated posts in various countries would now be trying to report to and take orders from Mussolini.

In the South there was now a 'Royal Italian' Government (at Salerno until Rome was taken nearly a year later). The Italian Diplomatic Service was always royalist by tradition. And now all the major ambassadors and ministers in countries where they could safely do so, declared for the King and put themselves at the disposal of the Allies. In Madrid the Royal Embassy handed over all its cyphering material for inspection to the Allied Representatives.

Meanwhile the Salerno Government, being subject to the control of Allied Force Headquarters for its communications, made a case first that it needed to communicate with Royal Italian posts abroad and subsequently that it should have cyphers (deposited if need be) in which to do so.

All the cyphers hitherto used by the Foreign Ministry in Rome had obviously to be considered as now equally useless to either side. The Germans in Rome had ample opportunity of ransacking the "Ufficio Cifra". We know that they did so. At the same time they must have known that Royal Embassies would deliver up their stock to us and that men who knew all this material were in Allied hands. It was a reasonable conclusion that neither we nor the Germans would ever be concerned with these codes and tables again.

It should be recorded that it was the unreasonable that happened. When cypher traffic was intercepted, as in the French case, both sides were found using the compromised cyphers.

Owing to the urgent needs of Japanese and other sections at Berkeley Street the Italian Section had, as soon as Italy surrendered, been quickly dissolved. Fortunately there remained a small nucleus to examine any cypher traffic which might need investigating, and its problem, for the first six months was (again contrary to all reason) not a Fascist, but a Royal, pro-allied cypher.

For the first seven months the Fascist traffic which was intercepted was almost entirely in plain text. This was probably an order made by the Germans, since, from an intercept in December we knew that they were reducing the strength and status of Fascist posts everywhere and they may well at this juncture not have had time to bother with supervising Fascist cyphers.

Traffic was intercepted between Salò and Berlin, Madrid, Tokyo and Budapest. Between September 1943 and January 1944 there were 800 intercepts, none of them separately important, but together giving a picture of the set-up at Salò the attitude of the Germans, and who were Mussolini's adherents abroad.

But there began to be intercepted in January cypher traffic between the Royal Embassy in Madrid and the Salerno Foreign Office which was entirely different from any known Italian Diplomatic Cypher. It was found to be a transposition, believed double since prolonged efforts failed to break it. We were in fact spending our efforts on a system which Allied Military headquarters had approved (and as we later learned) improved for the Italians on our side.

Eventually the key-book was sent to us. The keys at that time were derived from an English illustrated magazine. Though it was a 'deposited' system, it was not certain that the traffic was being read at Caserta [ie Allied Military HQ], and this was therefore done and the contents, when of interest, translated at Berkeley Street.

From this it became clear firstly that the system, being laborious, was unsuited to the long dispatches which diplomats have to send; secondly that the Italians were ignoring rules which were part of the security of the system, and thirdly that Madrid was being told to retransmit messages received in this system to a post which did not hold it. This last would be a major blunder, of which the Italians were quite capable.

But our interception of Royal Italian showed even worse to be going on, of which Allied Force Headquarters in Italy could not be aware, since they controlled only the use of cypher into and out of Italy. Between Madrid and seven other posts various of the old and completely compromised cyphers were being used.

This gradually led the Foreign Office to the conclusion that to advise the Italian Foreign Office and ensure that it used only secure cyphers and used them correctly a joint Anglo-American Mission staffed by Berkeley Street and ASA, Washington should be attached to the Allied Control Commission in Italy. This mission arrived in Rome on 1st October 1944. It consisted of the head of the Italian Section at Berkeley Street with one secretary and of an officer who had worked in the Italian Section at ASA, Washington. It worked in Rome as a subsection of the Communications Section of the Allied Control Commission until the end of hostilities in Italy in May 1945. Its terms of reference were:

1. To watch Italian diplomatic communications and their security.

2. To provide the Italian Ministry of Foreign Affairs and the Italian missions abroad with suitable means of communication.

3. To keep the Foreign Office, State Department and Allied Control Commission informed of the contents of Italian telegrams.

4. To keep GC&CS and the corresponding US organisation informed regarding Italian cyphers and lines of communication.

The Italian Foreign Office, now reestablished in Rome, had been informed through the Allied Control Commission of items 1 and 2, and no doubt suspected 3. Contact with its cypher bureau was established at once on the amicable footing that our control was a necessity while operations were impending. Each party was well aware that it was dealing with its professional opposite numbers who had obligations which would be veiled throughout the official dealings by a good-humoured reticence.’

 

Austerity in Post-War GCHQ - and Help from Friends

Austerity, belt-tightening, peace dividend: at various points during my career HMG found ways to make my life less simple, either by cutting GCHQ’s budget and therefore the resource we had available to put against a particular issue; or, more simply, by making sure I had a pay rise of less than the rate of inflation (as happened for each of the last ten years before I retired).

But much as I might have grumbled at the time (and since) I have to say that I never faced the conditions that GCHQ staff faced in the first few years after they left Bletchley Park in April 1946 to return to London: in fact to a former Bombe outstation at Eastcote, between Harrow and Ruislip in NE London.

Bletchley Park was too big for the much reduced Sigint organisation, and staff who had been with the organisation before the outbreak of war had been promised a return to London when peace came. However, not only was the London location not civil service premises in Broadway, nor anywhere else in SW1 (as returnees might have expected) for that matter, but the premises they were to occupy in Eastcote were just the same wartime TOBs – Temporary Office Buildings – many of them had spent years working in during the war.

The country was broke: there was no money for infrastructure other than to make do, and mend what could be repaired. Much of the damage caused by the 1940-42 blitz and the V-weapon attacks of 1944-45 remained unrepaired. The housing stock, which was already substandard in the 1930s, was in a much worse state after a decade of neglect. Not only that, but food rationing was still in place – in fact was even more severe than in wartime.

The winter of 1947/48 was exceptionally cold, so cold that transport was sufficiently disrupted to cause the breakdown of the national distribution network. And this caused more problems in the cities than in more rural areas, as food could not be brought from the areas in which it was produced to the areas in which it could not be produced. Rationing of potatoes — something that hadn't happened during the war — was introduced.

So I think of my predecessors in Eastcote in December 1949. It was more than ten years since the beginning of the war, and yet heating was limited, public transport was the only way to and from work, and salaries were low and highly taxed. (Income tax was 45%, while for those on salaries above £2000 (there were very few at GCHQ) the higher level of tax was 55%.) The 56 conditioned weekly working hours of wartime had been moved back to 42.5 (exceptionally: the rest of the Civil Service was on 45), but it had already been announced that these 42.5 hours would rise to 45.5 hours the following year, with Saturday morning working reintroduced. In short, things were bleak.

As I said at the beginning, this was austerity with a vengeance, but in December 1949, as had been the case in December 1947 and December 1948, there was a tiny challenge to the austerity in GCHQ. Not a staff rebellion, not some outrageous act of theft, but the action of a large number of unknown friends. On 22 December 1949 staff received a message from AD(P) (Assistant Director Personnel. responsible for HR, Finance and Personnel Security). This post was held by Eric Jones, who would later become GCHQ Director. He came from a family of textile manufacturers based in Manchester who had been commissioned into the RAF as a Group Captain and had been posted to Bletchley Park specifically to make Hut 3 (German Army and Air Force analysis and reporting) work efficiently and productively. For the time, his approach was singular: he dealt with all of his subordinates with courtesy and tact, treating them all, from the most junior, as individuals. This brought out the best in his staff and as a result Hut 3 tended to be a happy place to work. (Some of his Directorate colleagues referred to him, somewhat sniffily, as 'the Manchester Businessman': it was not meant to be complimentary.)

AD(P)'s message informed staff of the Christmas signal that the Director had sent to the Coordinator, the head of the (US) Armed Forces Security Agency (AFSA), the predecessor of NSA.

The Notice reads as follows

‘Signal to Coordinator

The distribution of gifts so kindly sent by staff of AFSA has just taken place. This wonderful generosity, and the friendship of which it is the token, is warmly appreciated by all, whether lucky or unlucky in the draw. By it, over the past three years, about a thousand of GCHQ staff, covering the lower salary scales, have shared the parcels.

Please be good enough to convey this message to the staff of AFSA, with the thanks and best wishes for Christmas and the New Year of the whole staff at GCHQ.’

What had happened was that the staff at AFSA, who had heard through US liaison staff in London of the hardships facing their UK opposite numbers, had decided in 1947 to send Christmas parcels to GCHQ. At GCHQ, it was agreed that only clerical staff, the people who couldn’t easily afford to supplement their rations on the black market, should enter a draw for the parcels. And the clerical staff quickly worked out that by pooling their tickets and sharing the results of the draw for the parcels, the chances of some supplement to the Christmas ration was more likely. In this context, ‘about a thousand of GCHQ staff, covering the lower salary scales’ meant just about everybody in clerical positions. It's hard to imagine nowadays, but each of those food parcels made a difference between an austerity Christmas and an old-fashioned "eat, drink and be merry" Christmas.

This is also a demonstration of the fact that UKUSA has always been something more than an arrangement of expedience. The relationship with partners has an added dimension because of the continual contact between individuals from each of the agencies. At this level, the relationship isn’t between allies so much as friends, and friends look out for each other. (And it shows the boundless generosity that characterises so many Americans. In November 1948 the organising committee at AFSA was quoted $1000 for the thousand eight ounce packages of tea and the thousand one pound packets of sugar it sent alongside the CARE packages.)

I wonder what the Ghost of Christmas Yet To Come would have had to show all those people: their dreams might not have gone much beyond the basics: enough to eat; a warm house; nice clothes; but imagine how pleased they would be if they could look into the future and see that as the world changed, and the country managed to sort itself out, the need for the food parcels went away.

One or two old lags might recognise this as a reworked version of something I posted internally in GCHQ when I was still ‘inside the wire’. My thanks to my successor for letting me retell the story and for letting me use the photographs.

Sigint in Fiction

I had an articled published last month in the John Buchan Journal (unsurprisingly, the journal of the John Buchan Society). It is about the way that John Buchan drew on his First World War experience as a customer of Sigint to use cryptanalysis in one of his books and in a short story to advance the narrative, and to develop characters.

A key point, I feel, is that it is impossible to describe the process of cryptanalysis in a work of fiction and make it interesting for the general reader. Buchan’s answer was to give a vague idea of an encryption system: Playfair in one, double transposition in the other; make the reader understand that this is something really difficult, and that therefore the practitioners have to be intelligent, and not just lucky; and move on, making the story (or the relevant part of the story) about people who break codes or ciphers, not about the process of breaking them.

Two of the novels of Dorothy L Sayers have cryptanalytic sub-plots. Lord Peter Wimsey certainly shows himself to be intelligent in breaking the messages concerned, but the eight or nine pages in Chapter 28 of Have His Carcase, in which he goes step by step through the process of breaking a message enciphered with Playfair, and takes no prisoners while doing so. (She doesn’t try quite as hard in The Nine Tailors, in which she describes in just two or three pages an encryption method which uses as its key the course of the treble bell in a peal of Kent Treble Bob Major.)

Perhaps the process of cryptanalysis is of limited interest to the reader of a novel because the process of Sigint – interception, analysis, cryptanalysis – is analogous (though less interesting) to having somebody physically steal a copy of the message. The point the author is aiming at is that somebody who shouldn’t have had access to the message now has it, so the story can move on. There are more and less interesting ways of doing this, and pausing the story so that Lord Peter can show off is at the ‘less’ end of the scale.

I’m not aware, and I think the reason why is clear, that there is that much at all about Sigint in literature. I’ve dealt with Buchan; From Russia With Love is the only Bond story which touches on it (the attempt to steal a Soviet encryption device draws from British plans to steal Enigmas during the Second World War that Ian Fleming was closely involved with); John Hale’s The Whistle Blower is about cryptologists, but not cryptology; are there any more?  (I'm speaking of British novels.)

There is one candidate for a novel about GCHQ, though: The Tin Men by Michael Frayn. His National Service was spent being trained in Russian so that he could be a Sigint linguist. The novel is about a research organisation: The William Morris Institute of Automation Research. The institute is at the forefront of automation: computers are being programmed to write newspaper articles, produce popular TV shows and say prayers. But problems begin when programmers begin to program computers to welcome HM The Queen on a royal visit to the Institute. No spoilers, but it is as funny a book as you would expect the author of Noises Off to have written, and could probably claim the title of the first satirical novel about AI.

The novel was published in 1965, at a time when any mention of GCHQ and the work it carried out would have brought an author to charges under the Official Secrets Act. But in the succeeding fifty years the story has been passed down in GCHQ that The Tin Men was Frayn’s attempt to make sense of the UK’s national Sigint organisation.

Do read it!

 

More Detail on Cairncross at Bletchley Park

Last week saw a major release to The National Archives of MI5 files, mainly concerning ‘The Cambridge Five’ or at least the three of them, Philby, Blunt and Cairncross, whom MI5 had the opportunity to investigate (Burgess and Maclean having skipped the jurisdiction).

I’ve obviously not been able to go through this new material in any detail, but I have looked at the parts of the Cairncross files which shed light on his time at Bletchley Park.

I covered some of this three years ago in Sigint Historian: Personnel Security at Bletchley Park - Part Two

There is some interesting new (to me) detail. Cairncross had been encouraged by his controller to join GC&S or SIS. When he first identified to MI5 the battle in which the Ultra material he had passed to his controller had given the Red Army a decisive advantage over the Germans he claimed it was the Battle of Kharkov (Kharkiv, of course, today), not Kursk. Interviewed in 1967, Peter Wright asked him whether he could remember anybody at Bletchley who might have been pro-Russian: his answer was that everybody at BP was pro-Russian because of the events of the war, but he named three people Hugo Gatti, Philip Pounsey and Douglas Parmier, but these are likely simply to be three names plucked out of the air. (None appear on the Bletchley Park Roll of Honour.)

Edit: I stand corrected. Hugh Gatty, Philip Pouncey and Douglas Parmeé are all on the BP Roll of Honour.

More insidiously, he said that his Soviet controller had once asked him why a former FO colleague, Roddy Greiffenhagen, had transferred to GC&CS, and whether this move would hit Greiffenhagen financially. After some investigation, MI5 discovered that Greiffenhagen was eased out of the FO ‘because of his total inability to do the work required’, and concluded that Cairncross was likely to have passed biographical details of other FO colleagues to the Soviets as well. He admitted to having passed a couple of pen pictures of GC&CS colleagues to his controller, but couldn’t remember who they were. This part of the release leaves a question mark over Greiffenhagen’s reputation: it’s a bit unfair, as there was no doubt an investigation subsequently: one, potentially, for a future release.

Cairncross said that his controller was annoyed when he engineered a transfer for himself from BP to SIS Section V at Ryder St to work on counter espionage (a section to which GC&CS deployed several members of staff as it began to resume its work on Soviet targets), the annoyance, he claimed, being because of the quality of the material he was passing from Hut 3. I wonder whether the presence there already of Kim Philby might have been a better reason for the controller’s annoyance.

There is a curious tale of Cairncross at Ryder St pestering somebody (whose name is redacted) at Berkeley St (where GC&CS diplomatic traffic was worked) to get information about a breakthrough made by Berkeley St. The unnamed person said that he had probably been indiscreet at the time, but reflecting on matters in 1967, thought that as Philby was handling this material, his indiscretions were probably not that relevant.

There is an amusing 1955 account by Henry Dryden (a Second World War Siginter who stayed on after the war). In the spring of 1949 Cairncross phoned Dryden at GCHQ, then at Eastcote, inviting him to lunch, and over the soup course asked whether GCHQ was a successful with Soviet encryption as it had been with German. Dryden ‘mumbled some sort of non-committal reply, bringing in the phrase “one-time pad” and tried to give the impression that one could not do anything to make progress’. This story, at least, is already in the public domain, having been told by Dryden in 1993, in a postscript to his contribution to Hinsley and Stripp’s Codebreakers.

There may well be more details in these files, and I’m sure there will certainly be in other files in the release, but three things stand out to me. The first is something I have mentioned before: how little the MI5 investigators knew of (never mind understood) how Sigint worked and their consequently appearing to deal with it as not as key an issue as were some others. The second, is that my judgement of three years ago, that of course it was likely there were other Soviet sympathisers at Bletchley, but that they were unlikely to have been able to pass on many great secrets, hasn’t changed, not least because of the third: that only if somebody had access to Bletchley’s serialised reported intelligence would it be possible to pass intelligible information to a hostile intelligence service.