forum How Do You Language???? A Step-By-Step Guide Part 1: Word Crafting
Started by @Oakiin
tune

people_alt 61 followers

@Oakiin

Ever sat down and thought to yourself…I wanna make my own language!
And so you open up Notebook or WorldAnvil, or Campfire, or simply grab a piece of paper or open a document…and then you blank out. Questions pile in, overwhelming you in details you have no idea about. What words do my cultures know? What do I do about phonetics? Grammar? What's sentence structure? How do I choose? How many words do I need to come up with?? Do I just grab Websters and start making up nonsense words?
And soon you're doing something else, because, let's face it, that's too much to sort through easily and not feel panicked and inconsistent. Well fear not! While I am by no means an expert, and I don't have all the answers (yet…), I have a system gifted to me by an older sister, and some strategies of my own. Also, this is literally a guide you can follow, you don't have to read it and then still be confused So let's make languages! (Feel free to drop your own tips and strategies in the thread too, let's compile our knowledge!)

Making a language, step 1:
Pick your two favorite languages. Works a little easier if they have the same symbols, but as long as you know the phonetics, you're good. Compile lists of words from both languages, then start cutting them in half and stitching them together. Or mix them however you please. This will lay down the foundation of what your language will sound like.
For my example, I'll use Romanian and German, both of which I have studied.
Romanian words: | German Words:
— | —
pisică | andere
furculiță | teuer
La revedere | Nacht

Stitch them together like this:

  • Andriscă
  • Furteuliță
  • La Nachedere

There ya go. You'll need 100 of these. I could continue to use the same words and mix them in different combos, but I recommend you get a diverse cast of words, maybe grab a dictionary or two and just grab random words. This will up the diversity of sounds in your language!

Making a language, step 2:
Now that you have your list of 100 words, (don't worry about definitions yet), you count syllables up. For example, how many words are one syllable, and how many are two, etc, until you have counted all the words. Say 50 of them are one syllable it means 50% of the words you make will be one syllable. From here on out, you're going to want to use a 10 sided die, or an rng, such as the lovely one over at Random.org

So for me, I have one with 3 syllables, one with 4, and one with 5. So in my language, I would probably have longer words. I wouldn't know unless I made the full list, really, but this gives me an idea.

Then you do the same thing for what letter or letter cluster (like ai or nd- anytime two vowels or two consonants are next to each other) comes first in the word, then the letter or cluster that comes in the middle, and the ones at the end. Using the above examples:

La Nachedere
Letter Clusters Break down into:

La - which I count because it's part off the word in my language, not a different word tacked on, despite the space.. That means it may crop up in many places, so I want to mark it down. I may end up using it as a modifier, like 'er' or 'ing' for our language.

N - First stand alone letter!

a - Second stand alone! You will want to count them all, even repeats.

ch - the only other letter cluster in the word. These two letters go together and make a specific sound when placed together. These are what you want to note down.

e - is here by herself! She doesn't go with any cluster at all, so we could her as a stand alone. From here down, is just single letters! make sure to mark them down.

d -

e -

r -

e - the last stand alone!

Furteuliță
Letter Clusters Break down into:

F - The F and the U are the only stand alones in this word. The rest are clustered! This word is a great example of different clusters you can have.

u -

rt - two consecutive consonants. They may not make their own sound when placed together, but I can decide if they do or not later.

eu - the vowel ones are especially important for most languages, as they dictate most of how the language is accented. More on that later.

ță - these two are both marked with special symbols, so when put together, they may change the pronunciation of the letters. Something I would decide later. I could also choose not to cluster these, but I've decided that for this language, they are a cluster.

Andriscă
Letter Clusters Break down into:

A - stand alone vowel! keep this baby on her own.

ndr - this one has three consonants, which you want to note down as well.

sc - same as the above, and you'll note I left out the ă. I don't see this becoming it's own letter cluster, the ă doesn't affect how the letters are presented. If I wanted, I always could choose to leave it on, but that makes things more complex.

ă - another stand alone, this is a good thing! Stand alones give you your word diversity!

If you have a word that is too short to split, don't worry about it. Just count it as the one. Ex: The 'La.' If it were it's own word, it would have single use, and not ever be mixed in the stirring pot, so to speak. Slap a definition on it, and call 'er a day.

Making a language, step 3:
So now you're here. Good job! The last thing to do, is figure out those percentages.
Mine, which are inaccurate, given I only have three words, are one percent for each, really. If I wanted to get picky, I could break down the letters themselves, which means I end up with:
E = 5 (occurrences)
A = 2
R = 3

We'll call that there, since that adds up to 10! Yours should be at least 100, (though with an RNG, you can work with any number you end up with), but this is close enough for me. I have a 50% probability of my word having an E in it, a 20% chance of an A, and a 30% chance of an R. Obviously, this is very inaccurate, since I didn't calculate ALL of my letters, and you may have 'ugly' percentages, like 43% or 17%. This is a good thing! Don't be scared of them.

Now assign those percents and letters number ranges. for example, E has a number range of 1-51, A is 51-70, R is 71-100. Now I can roll a 100 rng, and where the number lands, voila, that's my letter. If you have a letter that only comes in once in all 100 words, assign it a single number. For example, 'ț.' If I had added ț in there, I probably would have assigned it the number 1. Then, ONLY if the rng rolled a 1 would that letter crop up, making it a rare letter in your language, like X or Z for us. And yes, it's awesome to have multiple rare letters! I've got ț and F, I say the more the merrier.

So, the last step to word creation is to keep all this info in separate lists, and then roll for syllable count, start cluster, middle cluster(s, if applicable), and end cluster. As far as letter clusters go, I usually manually tweak some words (I use a coin to decide if i mess with it or leave it be), then roll for letter clusters and adjust the word as I please. Be careful with this, as you can throw off the sound and feel of your language, so make sure you test-drive your words! Read them out loud next to some words you feel really capture the vibe of your language, and make sure they don't feel entirely wrong.

(Bad) example of all this, with made up percentages.
Syllable count: 2
Start cluster/letter: F
Middle cluster(s)/letter: eu, ch,
End cluster: ță
Final word: Feuchță
Now notice the HUGE letter cluster at the end. I can, if I want, make a rule about that regarding my language's pronunciation, but I can also just say they don't affect each other's phonetics. This word is pronounced FEW-chto, by the way :)

And that is how you come up with words for your language! If you try this out, please let me know, I may have written some of it slightly incorrectly, I spent so much time editing and re-reading I may have a few inconsistencies or mistakes. If you have any questions, ask away, I'll do my best to answer.

And yes, this is but one tiny part of the huge project that is language creation. Next time, we'll look at crafting grammar structures and don't worry, it's easier than you might think.
I hope this helps <3

@Young-Dusty here you are :)

@Oakiin

Hey don't worry, most of it's just counting, and rolling rng's! You know me, I would NEVER condone a system with lots of math xDD

@Akmedrah

I feel like i need to plug all this into a java program, at least the sylibal counting and rng parts and see what happens. If it comes out good ill post it here no guarantees though.

@Akmedrah

so, I tried my best, and did some research and I have failed. It seems that sylibal recognition is something that is the topic of thesis papers for computer sciences rather than something that is a "practical application" and that is just for English. It gets infinitely more complex when you deviate from the 26 letters of the English alphabet to be able to 'mash' two languages together like that. Sorry for getting yall's hopes up. I am going to try and come up with a "mixer" of sorts that will mash the words up for you and then ask for the sylibal count of each to do the math. but classes just started so it may take me a while.