By Gordon Rugg and Gavin Taylor
The story so far: This is a series of articles about the D’Agapeyeff Cipher, a short ciphertext that’s never been cracked. It first appeared in Alexander D’Agapeyeff’s 1939 book Codes and Ciphers. It’s a good testbed for codebreakers who have to deal with short texts. In the previous episodes, we’ve listed the codes described in the book, plus the worked examples that D’Agapeyeff used. In today’s gripping episode, we focus on a typo that Gavin found in Codes and Ciphers. It’s more exciting than it sounds, though admittedly that isn’t the most difficult challenge in the world…
Opening question: Does anyone apart from sad pedants really care about typos, and if they do, is that just a sign that they really ought to get a life?
In fact, a lot of people care a great deal about typos, in a lot of different fields, and for a lot of very good reasons, most of them involving money and the law. They’re also a serious issue if you’re applying for a job. A high proportion of organisations will reject a job application because of a single typo in the covering letter.
So why do people care so much about typos, and what does Gavin’s discovery mean for anyone interested in the D’Agapeyeff Cipher?
We’ll start with job applications and recruitment.
Here’s a quote from a career site, talking about job applications and covering letters:
A top complaint with every manager and HR person in our survey noted: “I stop reading when I find spelling mistakes.”
That’s a common finding in surveys in this area. Organisations take spelling mistakes and typos very seriously. There are two main reasons for this.
One practical reason involves logistics. Imagine that you’re a manager dealing with applicants for a job in your organisation. Processing job applications takes time, and you may be dealing with several hundred applications for a single job. So you want a fast, simple way of weeding out the non-starters, and typos in the application are one of those fast, simple ways. They have the added advantages of being objective (so you can’t be legally accused of subjective choices that can easily shade into prejudice) and of having a practical job-related justification. That’s where the second reason comes in, as the justification for using typos as a criterion for rejection.
Companies don’t like needlessly losing money, and typos can be a surprisingly effective way of losing money. Here’s a recent example.
A hotel offered a room in Venice for one euro cent instead of 150 euros. It was a misprint, but the company had to honour its offer. The cost to the company was estimated at 90,000 euros.
You might think that this was a one-off by some employee in a local hotel, but this sort of thing happens to major international organisations too. Here’s another example – pocket computers mistakenly on sale on Amazon for £7 each, instead of £192.
Neither of those broke the bank, though they weren’t exactly cheap, and the companies involved wouldn’t love the employee who made the mistake.
Here’s a more impressive example, where the error cost the company about $220 million:
The Japanese government has ordered an inquiry after stock market trading in a newly-listed company was thrown into chaos by a broker’s typing error.
Shares in J-Com fell to below their issue price after the broker at Mizuho Securities tried to sell 610,000 shares at 1 yen (0.47 pence; 0.8 cents) each.
They had meant to sell one share for 610,000 yen (£2,893; $5,065). http://news.bbc.co.uk/1/hi/business/4512962.stm
So, there are very good reasons why organisations are wary about taking on employees who might be responsible for mistakes involving small but important typos. You might be wondering about the laws relating to discrimination with regard to conditions such as dyslexia; that’s a good point, which goes outside the scope of this article. Dyslexia is a complex issue, which Gordon discusses in some depth in his book Blind Spot. We’ll leave this topic for now, and return to the D’Agapeyeff Cipher.
The conclusions from the previous examples are that some typos are minor mistakes that don’t really change anything, but that other typos are significant. If you’re working with codes, then typos can be a very big issue. Some codes are fairly tolerant of errors, but with others, a single mistake in the process can mess up everything which follows, leaving the rest of the message unreadable.
There have been suspicions over the years that this might have happened with the D’Agapeyeff Cipher. D’Agapeyeff was an amateur cryptographer, not a professional. It’s plausible that he might have made a simple error part-way through the encoding of his cipher, and inadvertently turned it into gibberish.
In principle, this shouldn’t have happened in a professionally produced book, with copy editors and proof readers involved. (This is another reason that most organisations are twitchy about spelling mistakes and typos in applications; if you’re putting out corporate written material, then professional copy editors and proof readers will be involved, costing time and money; they don’t love people who send them copy with a lot of mistakes in it, since the expense of correcting those mistakes will cut into the profit margin.)
In the case of Blind Spot, for instance, the copy editor and proof reader checked everything, including the gibberish text that Gordon had produced to demonstrate a point about the Voynich Manuscript.
In practice, though, there’s one significant point about the D’Agapeyeff Cipher which makes it different from the text in the rest of the book where it’s printed. The Cipher was a set of numbers, and the job of copy editor and proof reader would be to check that the numbers printed in the book were exactly the numbers that D’Agapeyeff said they ought to be. Whether or not there was an error in the cipher behind those numbers was D’Agapeyeff’s problem.
That’s where the typo that Gavin discovered comes into the story. It’s not a simple typesetter’s error or spelling mistake. It’s one that goes deeper than that. Here’s the relevant section of text.
It’s not the most immediately obvious of mistakes. However, it’s the sort of mistake that could mess up an encipherment.
In the image below, we’ve shown the error in red highlight.
The list of letters according to Valerio includes the final set of letters “UCX”. However, in the next line, D’Agapeyeff states that several letters are not met with in Valerio’s list, and includes “C” as one of the letters not in the list.
That’s inconsistent, and it’s the sort of inconsistency that could seriously mess up an encipherment if D’Agapeyeff made a similar mistake in the detail of his encoding.
So, in conclusion, this adds some weight to the suggestion that D’Agapeyeff may have made a mistake in encoding the D’Agapeyeff Cipher, and that the text in the Cipher may be hopelessly corrupt.
In the next episode we’ll discuss this possibility in more depth, and consider some suggestions about what the text of the Cipher might be.
Gordon’s book Blind Spot: