“To whom should this be? M,O,A, I…let me see, let me see. And at the end – what should that alphabetical position portend? ‘A’ should follow, but ‘O ‘ does….“
That’s Malvolio in Shakespeare’s Twelfth Night, trying to decipher a secret message he’s just picked up. He’s being pranked, but your average codebreaker’s work is deadly serious . . . especially in wartime.
America hadn’t yet been drawn into the Second World War in 1940, but it was obvious to our intelligence services that it was just a matter of time. We needed codebreakers – lots of them, and fast. But where to find them? The draft was already vacuuming up men of fighting age for combat duty, and the manpower shortage would only get worse. Hmmm…. How about women? Naval recruiters started canvassing elite women’s colleges in New England for exceptional math and language majors of solid character. The Army followed suit, focusing on teachers colleges in the South and Midwest.
By late 1941, there were 181 women at Arlington Hall, the Army’s cryptographic facility outside Washington D.C. After Pearl Harbor, that number exploded: 8,000 were at work decrypting Axis message traffic there by war’s end. And in February 1943, a unit of just two souls was added to that mass. The code they were working on was Russian.
Wait a minute. Wasn’t the Soviet Union our ally in that war? It was. But it had been Germany’s ally before that country invaded them in 1941, and if the two decided to make a separate peace, the U.S. and Britain would be facing the Axis alone. We needed to know, and to plan accordingly. So two codebreakers became six became dozens; new recruits and old blades were funnelled in, and by 1945 the unit was processing 400 encrypted Russian messages per day.
Into that top-secret whirlwind stepped an Indiana County hairdresser….
Angeline Rose Nanni was born in 1918 to Biagio and Philomena Nanni of Creekside, the third of six children. Angie’s knack for numbers was apparent almost from the start: she graduated early from eighth grade, but since she was too young to attend Indiana High School yet, she helped out with bookkeeping and delivery preparation at the family grocery on Arch Street while she waited. That same talent got her an after-school job at G.C. Murphy’s in Indiana once she was allowed to enroll. Half of her 1937 graduating class would be in uniform five years later, but Angie was destined for even greater contributions.
After high school, Angie left Creekside to work in Harrisburg. Returning a year later, she joined her sisters in beauty school and then in their Blairsville and Indiana salons, where she kept the books and left the styling to Virginia, Mary and Jennie. The Nanni sisters closed shop “for the duration” in 1944 and headed to Washington to pick up war work; Virginia joined the Marines, then returned with her sisters to their salons after the war. Only Angie stayed on in D.C.
In the autumn of 1945, Angie was invited to test for a job with SIGINT, the Army’s Signal Intelligence Service. She assumed it would be to qualify for routine clerical work, but when she solved a cryptic math problem using only what is now called the cognitive unconscious, SIGINT knew they’d found their next Russian codebreaker. The other test takers – all college graduates – were sent home.
The Russian unit, codenamed Jade at first but Venona in later years, was kept secret from other codebreakers at Arlington Hall. They sat in a screened-off area where none but those in the unit were allowed, and each evening they locked anything with Cyrillic (Russian) characters on it away with their codework. They were never to speak of their work to anyone, not even each other, once they crossed the threshold. Family and friends assumed they were clerical drones, and the “Code Girls” let them think so.
Venona was divided into Traffic, Reading and Back Room sections. Angie started in Traffic, sorting encrypted messages by source. But project head Gene Grabeel, wanting to make best use of Angie’s gift for pattern analysis, soon moved her into Reading. There she typed the most promising messages into an IBM keypunch machine and analyzed the resulting tables for character-sets that implied the sender’s use of a certain “duplicate pad” for encryption. That pad was the entire reason for Venona’s success against an otherwise unbreakable Soviet code.
What was it, and how did we know about it? The key to the code’s invulnerability was its use of an ever-changing set of character-substitution sheets in the coding process. Compiled in pads, each was to be used just once, then destroyed. But when Germany’s 1941 invasion threatened the plant where those pads were produced, its presses, paper and personnel were moved deep into the Russian interior. Until new sheets could be configured, it was necessary to reprint existing ones. These “duplicate pads” (a fraction of 1942’s total output) were dispersed to Soviet spy-centers around the world in an effort to avoid the very pattern-detection Venona would eventually accomplish. Most were used up by 1945, but that was enough: the code was cracked in 1946.
It didn’t hurt that the Finns (intentionally) and Japanese (unintentionally) gave us a hand. Finland had been invaded by the Soviets in 1939; resistance fighters captured a partially-burned coding pad, which they sold us in 1944. They’d also identified structural clues embedded in Soviet message traffic, and that knowledge was sold to Japan. When it turned up at Arlington Hall in decoded Japanese transmissions, it was passed on to Venona. There, geniuses like Meredith Gardner and Richard Hallock broke the duplicate-pad coding . . . yet without the intuitive scanning and labor-intensive matching done by the likes of Angeline Nanni, Venona could never have uncovered the astonishing revelations that followed. Rosie the Riveter had nothing on Connie the Codebreaker!
Once the decrypted messages were read, it was plain we’d gotten more than we bargained for. No news of Russian-German negotiations, but a stunning parade of code-named Soviet spies in the US, UK and Australia. Hooking up with the FBI (and later CIA), Army SIGINT was able to match details of actual persons with those mentioned in Venona intercepts. And what a list! In the 3,000 duplicate-pad intercepts decoded by Venona, 349 spies were mentioned, half of which were ultimately identified. To name just a few:
- Henry White, Assistant Secretary of the Treasury
- Klaus Fuchs, Manhatten Project nuclear physicist
- Lauchlin Currie, White House advisor
- Kim Philby, British intelligence liaison
- Alger Hiss, UN and State Department officer
- Julius Rosenberg, Army Signal Corps engineer
- Duncan Lee, OSS officer, one of 20 identified
Several were tried, some turned state’s evidence and a few escaped to Russia. Ah, and then there was the one in Venona itself. Codenamed “Link”, he’d even been mentioned once the intercepts but was not identified until 1950 as linguist William Weisband.
Though only a small percentage of the 1942–48 intercepts were deciphered, Venona’s impact on history was remarkable. Angeline Nanni and two others stayed on to the end, packing away the material for storage in October 1980. Until that material was declassified in 1995, the public remained ignorant of Venona and what it revealed about the extent of Soviet penetration into American government, military, academia and journalism.
Angeline Nanni is still with us, one hundred years young and living in Washington. There are still Nannis in Creekside, too, as you’ll see if you visit the family restaurant there. As for their Angie’s place in history, well, they could surely be forgiven for saying (with apologies to Malvolio):
TPNFB SFCPS OHSFB UTPNF BDIJF WFHSF BUOFT
TBOET PNFIB WFHSF BUOFT TUISV TUVQP OUIFN
(Can you crack this “simple substitution cipher”? we’ll post the answer next month on Facebook)