# Designing and Implementing a Crypto-Puzzle Language

Currently I am in the planning stages for when I plan to run a game of D&D 4E, but this question its self is more system-agnostic.

I want to design a written language that will come up in some of the puzzles I present to the players. Some of these puzzles may be alongside other similar puzzles, while some would be the main focus. The main thing is I want to use a persistent language, and not just random symbols each time.

Ignoring coming up with such a puzzle language(a task for another time), I am in need of ideas on how to present such a puzzle, but not make it break future puzzles involving the same language(aka a meta-gamer bringing his own decoded language reference which is more than the characters have gathered in game). While I don't want to limit creative thinking on these puzzles, I just need ideas to keep them from becoming stale.

--UPDATE--

I feel the wrong question is being answered. How to design a secret language is not what I need help with, but that is what is being answered.

What I need ideas for is implementing a language as part of various puzzles, where success or failure at a previous puzzle could affect the outcomes of future puzzles involving the language. I don't want to make it to easy to early, but at the same time I don't want making a single failed puzzle fail the whole lineup.

-
Why not use a real life script like Korean but write as if you were using English? – Sardathrion Jul 1 '13 at 6:38
Point of order, Comprehend Language, a 1st level ritual, completely obsoletes this aspect of the challenge. How do you plan on dealing with this? – Brian Ballsun-Stanton Jul 1 '13 at 6:53
Short of restricting use of that ritual, I have not thought of that yet. At my current point in planning, I am compiling various ideas I have to filter and narrow down latter. Now that I am aware of this possible work around, I can work construct ideas around it. – RMDan Jul 1 '13 at 7:09
So is your real question something like "How to design a series of puzzles so that each one gives clues for the next one, but so that failing a single puzzle doesn't make all the rest impossible?" – Ilmari Karonen Jul 1 '13 at 7:59
Could you clarify on language? Even after the update I'm not sure where to focus on an answer. You could use a different alphabet with unfamiliar symbols (bonus points if you use ones with phonemes the players aren't familiar with). You could invent a vocabulary. You could be talking about a new grammar with different syntax rules (ie, ordering by verb-subject-object. Steve kicked the ball would be kicked Steve the ball.). Or this could be about different waits to encrypt a message. – valadil Jul 1 '13 at 12:57

This is a question of character knowledge versus player knowledge.

Cryptanalysis is a fun, if little used, human skill. Some players will delight in solving cryptographic puzzles, and 4e certainly doesn't prohibit the gamist approach of "challenge the player."

On the other hand, cryptanalysis is highly binary. People either really like it, or basically can't stand it, especially if you start branching out into some of the more interesting, but trivially solved cyphers (playfair, Vigenère).

I, unfortunately, had to learn this fact the hard way. Players, as a rule, will take one of two approaches to puzzles. The first approach I learned when I gave a cryptogram in a sci-fi game: "I hire an expert from town to solve it for me." is the general pattern here. To make this approach interesting, the cryptanalysis must have extrinsic complications such that one can't submit it to an analyst (magical or otherwise) and wait it out. Beyond that, having a player go "and I solve it. what does it say?" is absolutely valid, and should require either a pure int check, or a skill check of an appropriate skill the player is bringing to bear. I would personally recommend looking at the serious skills series to see how to make this approach more interesting, as players can add depth to their character by describing how a particular facet of knowledge was brought into play. Still, if you're not expecting it, it comes as a horrible horrible surprise. Some people just don't find puzzles fun.

For the players who find puzzles fun, especially for a 4e game, take a page from Portal.

Teach players one component at a time of the puzzle. It'll seem way too easy to you, but you know the context. For a crypto puzzle, start by teaching the alphabet. (Maybe touch the "true characters" not to get shocked) In subsequent challenges, teach the order of the alphabet. Again, always give an example of what you're looking for before you ask for what you're looking for. (Or, provide a way for rapid abductive reasoning without penalty.)

This way, after quite a few sessions, your players will be able to manipulate the language as a tool to solve problems, and may occasionally be assisted by their characters-as-characters.

An interesting touch would be to make the language three dimensional, and have the puzzles follow the rules of the game Zendo. Then the rough series of questions can be "complete this sentence." or "find the incorrect thing." Both are completely appropriate to solving a crypto-language, and both can engage players, instead of abstracting the puzzle to their characters.

-
This is great. I like the Portal aspect. Also the Zendo game adds another layer of puzzles I could use. – RMDan Jul 1 '13 at 16:40

It is very easy to create a 1 to 1 cipher which will sound linguistically plausible. I've done this several times, for many different languages. It's also easy enough to break that, given enough text, the players could figure it out.

Effectively, what you do is map vowels to other vowels, and consonants to other consonants, and drop Y (and likely also C and Q). Another way to do this is to map sounds to sounds; this turns out much more realistically than alphabetic substitutions, but it's much harder for players to break. I'll give an (admittedly poor) example of the first one here:

A -> E
E -> O
I -> A
O -> U
U -> I


Then, map consonants to other consonants in a realistic, but creative way. The letters you map will determine how the "language" sounds. Take a look at this chart when creating the language; if you want hard sounds, for instance, you might map K to S.

Thus, create the table, and remember to skip the letter Y. You may also want to drop the letters C and Q; C's sounds are covered by both K and S, and Q's sound is covered by KW:

B -> L
C -> H
D -> K
F -> J
G -> X
H -> F
J -> N
K -> M
L -> D
M -> V
N -> W
P -> C
Q -> T
R -> Z
S -> B
T -> S
V -> P
W -> Q
X -> G
Z -> R


You're effectively creating a cryptogram puzzle at this point. So, let's take a phrase (You'll note that I interpreted Y as I):

Hello, my name is Richard.

Foddu, va wevo ab Zahfezk.

Obviously, you're going to encounter some conflicts. Simply put, Zahfezk really doesn't roll off the tongue. You may want to stick with that, though. It's okay to tweak it a bit at this point - but remember, if you want them to figure it out, don't change too much.

Note that this cipher probably won't work for most things, since I didn't give it all that much thought when creating it, and I didn't drop C or Q.

Now, it's very easy to create a symbol set for this. Depending on how in-depth you want to go, and how hard you want your players to work, you can either do this with another 1 to 1 cipher, or you can create a phonetic transcription. Either way, it's a simple as creating a new symbology, either for sounds, or for characters.

Note: Never translate names directly. They will not come out right, I guarantee.

# Playing around with phonemes to specifically customize your language

If you do choose to create a phonetically structured language, you can much more easily create a specific custom language. Feel free to play around with digraphs and trigraphs; your language no longer needs to map perfectly both ways. For instance, if you wanted a more snake-like language, you might map K to SS; if you wanted a Nordic-sounding language, you might map B to SVE (as in Sverdya).

You can even mess around with vowels, though be aware that you'll screw up a lot or words if you do this. A draconic language may map O to KOL (as in D'kolbochya, the Great Palace of Dragons).

# Presenting the language puzzle

All languages evolve over time. In order to keep things interesting, you're going to need to keep making symbols. The core structure of the language is already there. While we don't see it so much with English, it's fully possible for a language to accrue more symbols.

Start with complex symbols. Make them intricate and difficult to copy. (Optional: don't let your players keep notes on the language; I wouldn't do it, but it would alleviate the problem.) Then, as you give them more and more texts, slowly change out the symbols for simpler, cleaner ones. This has the added benefit of giving the language an evolution path; it becomes much more realistic when the language changes.

Lastly don't let your puzzles rely on one immensely complex aspect, like a written ancient language. If your players only need to break the cipher to recover the ancient crystal key, then they are going to smash your language to pieces. Instead, base puzzles on multiple aspects; leave language-deciphering as a pastime for the players. You don't want to alienate them with a frustrating puzzle. Your group may become dependent upon one person to decipher the text, if they break it. Additionally, a player with Comprehend Languages could break the puzzle entirely, without effort.

Use language puzzles in moderation.

-
While very informative, this did not cover the question asked. I asked for ideas on putting a language like this to use in game as part of many puzzles. This will help when creating the language I will use, but is not a complete answer. – RMDan Jul 1 '13 at 6:37
@RMd Working on it, sorry... Ran out of time and had to leave for a bit – Emrakul Jul 1 '13 at 6:43
@RMDan More words than I thought it would be, but there it is! – Emrakul Jul 1 '13 at 6:55
Thanks for adding more. I will still wait to see if more people answer before deciding to select the top answer. – RMDan Jul 1 '13 at 7:07