Rick's b.log - 2022/07/24
It is the 21st of November 2024 You are 3.135.214.175, pleased to meet you!

Decoding Paper Girls

Now, the comic came out about six years ago, back in 2016. So it actually predates Stranger Things, not that this will stop people saying it's a rip-off (just as Ghost In The Shell was accused of ripping off The Matrix by the clueless). This means, I'm sure there are others who have already done this work and sorted out what the symbols mean.
But where's the fun in that? ☺

Paper Girls, the other language.
© Brian K. Vaughan and Cliff Chiang.

Immediately, two things struck me. The symbols did not appear to be random, and there were spaces between runs of symbols as if indicating word boundaries. And that many of these runs of glyphs ended with , which would indicate some form of punctuation. Perhaps an exclamation mark? The activities didn't suggest a question was being asked.

It was, however, undecodable. At least, it was.

Then I got to this panel.

Paper Girls, a four letter word.
© Brian K. Vaughan and Cliff Chiang.

So, Bedouin bloke looks up, sees something he really doesn't want to, and utters a four letter word. Four letters if one assumes that   is some sort of end-of-sentence-punctuation.

Well,   being a four letter word provides me with two further pieces of information. It's quite likely to be a substitution cipher, that is to say, these symbols are standing in for alphabet letters. And the second? Well, what language has two well known four letter exclamations?

So I then worked my way back to the utterance at the top of the first image. The one that goes   because this ended with the same two characters (and punctuation).

Using a word finder, I searched for words with exactly nine letters using the key "???????CK".
There were 157 matches, none of which looked like that made any sense in context. But, look, just prior are double letters. Which means I need to look for any match that has a doubled letter, then end in "CK".
There were no matches.

So then I tried "???????IT". 39 matches, and using the double-letter test, it was whittled down to three possibilities:

• goddammit
• presummit
• whodunnit

This is where it gets a bit hard. Since I feel I can now have some level of certainty that   is "shit", then it is possible to strike off "whodunnit" (second letter mismatch), and "presummit" (fourth letter mismatch).
There was, of course, the slight problem that "goddammit" didn't seem to fit either that the third and fourth characters were different... although, they were mirror images of each other. That seemed a bit odd.

But, honestly, I'm out of ideas, so let's just go with "goddammit" and see what happens. Either it'll work, or it'll fail.

That means I can make up a substitution chart that looks like this.

First attempt at decoding.

Back to the first image, there's a larger section of text. Applying the knowledge from the above chart, the text becomes "S??IT ??! ???? M??T AT TH? S??O?D ?O??I??!".

There's a lot in there that is currently unknown, but there's something else we can learn. A three letter word that begins with "TH". There are only three. "The", "Tho", and "Thy", and it's only "The" that is in common use. Which implies that   is the letter E.
If this is so, looking at the text again gives "S??IT ??! ?E?? MEET AT THE SE?O?D ?O??I??!".

Now there's an answer for why the 'd' in "goddammit" was weird. It seems that if there's a doubled letter, the symbol for the second one is a mirror image of the first. This was determined by "MEET", which is the only word that seems to fit into context of things ("me?t" gives meat, meet, melt, and ment). Since the 'E' symbol is following the same inverted pattern of the 'D' in "goddammit", there is some merit to this.

"SE?O?D" can only be one thing: "SECOND". Which means I can add a "C" and and "N" to the known symbols.

What's the first word of the second sentence? It is "?E??" which gives 710 possible words, but there's a double letter there. So... Bell, cell, cess, dell, fell, fess, fett, jeff, jess, kell, less, mell, mess, ness, nett, pell, repp, sell, serr, sess, sett, teff, tegg, tell, terr, vell, well, yegg, yell, and yett.
As you can see, quite a few can be discounted because I already know some of the letters, so know which ones they aren't.

"Bell meet at the second...?"
"Fell meet at the second...?"
"Jeff meet at the second...?"
"Kell meet at the second...?"
"Pell meet at the second...?"
"Repp meet at the second...?"
"Vell meet at the second...?"
"Well meet at the second...?"
"Yegg meet at the second...?"
"Yett meet at the second...?"

Some of those can be ditched. Yegg makes no sense. Neither does bell. But nothing stood out. I looked at it awhile.
Nope.

So I went and put the kettle on.

Just as I was pouring freshly boiled water over my little bag of Tetley, it came to me. It's not "well" as in doing well, it's "we'll" just without the apostrophe.

"We'll meet at the second...?"
Now that makes sense.

Isn't tea great?

Now that 'W' and 'L' are known, they can be plugged into the text, to give "S?LIT ??! WE'LL MEET AT THE SECOND ?OL?I??!".
The first sentence, second letter first word, and second letter second word are the same. So I'm going to go out on a limb here and guess "split up".

I can't do anything with "?OL?I??" as there are too many possible matches. But, for now, the substitution chart now looks like this.

Better attempt at decoding.

As I translate more, and get more words, the chart will expand. This is as far as I have done for now. Phew!

The chart and symbols were recreated using DrawPlus, using an A3 size page. 4pt line width, and 128pt text.
The vector file was converted to an anti-aliased bitmap using InterGIF. Paint was used in case minor adjustments were necessary.
Scaling and converting to JPEG was performed using ChangeFSI. A touch of pre-sharpening was added to provide better contrast. The in-line text was scaled 1:7, and the charts were scaled 1:5. I'm thinking 1:4 might have been better, but the chart needs tidying up and spacing out. I'll do that once I have all the letters sorted.
This document, as most of my blog articles, was written in Zap, and it along with the drawings, were uploaded using NetSurf.
All of this was done using RISC OS 5.29 on a Raspberry Pi 1B running off of a big battery pack (15,000mAh), and using a 7" 1024×600 display panel. Two hours and forty minutes of uptime, and it has used 27% of the charge.
Photos taken using a Xaoimi Mi 10T mobile phone, cropped in the built-in photo editor, and uploaded directly from the phone using the Android version of Firefox.

Now to sit out, enjoy the nice weather, and... I don't know. Something. But whatever it is, it must surely involve at least another cup of tea!

Your comments:

Rick, 24th July 2022, 18:54

What I've just realised looking at this again (in between watching stuff on Netflix) is that "?OL?I??" is actually "?OLDING".  I had enough letters to make out all but the first. So, Either Folding or Holding? It isn't Bolding or Molding as I know what B and M look like.    Honestly, "We'll meet at the second folding" and "We'll meet at the second holding" make about as much sense as each other.    More decoding needed, but not today.

David Pilling, 25th July 2022, 03:58

Good work on the logographic language.

J.G.Harston, 25th July 2022, 10:05

On reading the first paragraph my immediate thought was "that's just Chinese, silly...." but then noticed there was too small a set of characters, and wondering how they were using a "100 basic words" type text.    Well done. :)

Rick, 25th July 2022, 11:25

I was mostly working on the assumption that it wasn't going to be an actual different language, just different writing - as it would be a bit much to have them speaking Chinese... Would really narrow down those able to understand it. The text was clearly *something* rather than random gibberish, and *something* is intended to be decoded. ;)    I wonder how they'll handle this in the TV adaptation.