April 1st, 2011. Roman Numerals and the Somerton Man

When I'm not playing guitar I like to spend long hours trying to solve mysteries from 1948.

I'll assume you're familiar with the story if you've made it here. If not, take a look at the links on the left first.

My solution to the code in the Rubaiyat poetry book is that it's a phone number.

The number is encoded using a mostly random string of 44 characters. The actual phone number is just the Roman Numerals. The rest of it is very effective distraction and if you ignore it you get:

MMMM D C L VIII = 4658

Variations:

Maybe the V on the last line is actually an I, but I doubt it, because that gives us 4 I's which is a not ideal (but allowable according to some rules).

We might also allow the Ws to be Ms, as the lettering is ambiguous.

This would give us: MMMMMM D C L VIII = 6658.

Or we might use a notation where the W = 2 Ms, I have seem this notation used at times: MMMMMMMMD C L VIII = 8658

A further possibility is that the first 2 lines and the crossed out line of the code, are a separate code to the bottom half, starting below the arrow lines.

The bottom half gives us: MMM C L V II = 3157

Area Codes in South Australia 1948

The X floating in the middle of the code is the area code for Glenelg, South Australia in 1948, where our story occurred.

It is floating by itself, because X is a Roman Numeral digit, and it would interfere with the phone number if allowed to sit with the rest of the text.

Roman Numerals and Probability

I had some heavy duty math guys calculate the probability that a random string of 44 characters will have a Roman Numeral encoded value that does not violate the rules for Roman Numerals. It came out at about 10%.

That's not quite improbable enough to make a strong statement, but it does lend some credence to the idea. If I was wrong, we' would expect 90% of the time that my theory could be defeated by saying "the number is not valid". So I'm either wrong and 90% unlucky, or I'm right.

The Crossed Out Line

In the image of the code we see the second line is crossed out, and partially repeated later on. This suggests that the code is not order dependent. It doesn't matter which order the characters are written out, because the Roman Numeral system will dictate the ordering of the signal component. There is not 100% true as sometimes the Roman Numeral subtractive rule creates ambiguities. For practical purposes it is true enough.

Contrived Patterns

The rest of the code looks a bit "desperately random", to quote Hannibal Lecter. Try writing out a list of random characters. You always end up writing familiar letter groupings like: AB, ST, TT, GO, AN etc... A complex cipher would produce a much more convincing stream of randomness.

Effective

No one can doubt the idea is effective. Simple to encode and decode, and no-one has thought of it all this time. Simple is good !

Not a New Idea - Chronograms

The idea of hiding information with Roman Numerals has been around for a long time. More often it is used as a Chronogram, which hides the year of an item's creation in a poem or similar. Often they appear in an inscription on a building. Run a Google search on Chronogram for more info.

Want to Learn More ?

If you get absorbed by this story, like many have, you really have to get hold of Gerry Feltus's book "The Unknown Man". See the link on the left and go buy a copy from his website. It's a great piece of work and gives you context of the era.

The Facebook group is a good place to start reading too.