Simplified Alphabet

This is pretty unusual and probably mostly along Sumer's line of knowledge, but why not.
I've taken a stab at this before, but a pretty lame one. Basically, I RP the Vaultian languages/dialects as being closely based on English, but evolved from there into several forms, mostly as simplifications.

As part of that I want to make an alternate alphabet, or, best of all, two alphabets, a short one used for practical fast typing and a complete one. Ideally the alphabets would have 16 and 31 letters respectively. The complete alphabet would have letters mapped to consonants and vowels as directly as reasonable, but without making the words overly clunky, so the question there is what to add.

The short alphabet is clearly more difficult. It should be able to represent the entirety of English vocabulary with minimal ambiguity and while being a subset of the full alphabet. That clearly has to involve more alternate pronunciations, but it has to be done such that words written in it are still recognizable by a normal English reader.
Any advice on what to cut and how to do it?

The 31 letter alphabet being some sort of longer short hand?

http://web.archive.org/web/20080209121816/...st100Words.html
I suggest taking the first few words from this list and turning them into a mashed up letter.


In Spanish, the letters k and w appear only in loan words and I wonder if I can make this post by not using the letter K in a word. X is not often used.

No, it being an alphabet.
Greek has 24, Latin has 26, Cyrillic has 33, Thai has 65, formal Vaultian has 31 or 32 letters.

Informal Vaultian, if it's possible to create, should use 16 letters, and be usable without explicit spaces. Translating 1-to-1 to the formal language and entered directly on stenotypes and now stenotype-like keyboards.

The current Vaultian language is approximately as related to Contemporary English as Early Modern English is to Middle English. I.e. it represents an organized amalgamation of English and fully integrated loanwords from other Indo-European languages that has been developing in its current form for about a century and has evolved some commonality.

Well if I may make some suggestions.

Im not sure if this will help with alphabet development, but whatever

Eliminate silent letters in word forms, do like the Hawaii thing where every letter present makes a sound(or contributes to the dipthong).

Eliminate vowels modifying each other---
short vowels aeiou stand alone, then for getting a long vowel sound create a specific letter(being an exception to the silent letter rule, all good rules need an exception) that is specifically a vowel modifier.
so your long vowel sounds (assuming the vowel modifer is ^)
Tha ste^d sta^d in tha shed. If that makes any sense.

Also perhaps merging ch/sh sounds or giving ch and sh sounds their own letter hence no reason to combine two letters in the word why shoe can be *oe and chew can be #ew ...again if that makes any sense.

Sorry I know Im not sumer, and likely these suggestions are rather remedial.

What I would do is assemble a set of all the vowel and consonant sounds in the language. Then have a character for each one. This would form the basis of the short alphabet. If it's too long, then you can collapse multiple pronunciations into single letters. For the long alphabet, find sets of characters that often occur together, like 'tion' 'th' 'ough' 'ph' 'gu' 'ively' 'izable', and give them their own letters.

This is mostly due to personal preference.
Edit: What I mean is that which letters you might choose to make would be depending on what character you'd want to give the language. If you just wanted it to be efficient you would chose the most common ones in english, or you could give it a Greek spin and add in ph, th, eu, or whatever.

Note that this is just something random I thought of, I don't have any schooling in linguistics.

I haven't formally studied linguistics (yet) so I'd be interested to see what Sumer has to say about this, but after a very quick mental run-through I've got some ideas of letters you could add to make 31.

Ö and Ä are the first to spring to mind and seem to me to be the most useful. You could also include an æ and the o version of that but that's pretty picky. I don't know much about consonants in IPA, which is what I'm using for comparison, so that's all I can really suggest.

I was going to give ideas for the reduced alphabet too, but my keyboard is half-broken making typing slow and the first idea I came up with (dropping the y and using u in its place) was confusing enough. So until I get ideas and/or a new keyboard that's all I have.
Although I think it should be easy enough to do away with c (ch can be represented by q) and replace it with k. I don't think doing away with h is a good idea but if you're going for compactness you could replace sh and z with x. IDRK though.

What are you trying to do?

I am a little confused. Are you trying to greatly simplify a writing system? Effectively create a new writing system much simpler then the existing? Or reduce the phonetics of the spoken language itself?

Edit: If it's just taking English as it is (Latin alphabet) and shrinking it to fewer written letters, but keeping the ones we have. That's easy as fuck. You could cut at least a few just by going over the sounds S/C/G/K, and the vowels.

No, absolutely no diacritics. I think they're unnecessary, and even if I didn't, every single letter must be unambiguously displayed by the smallest dot matrix possible (perhaps 4x6 for full alphabet and 3x5 for simplified).
For new graphemes I could draw on Greek or Cyrillic, or just make them up.

Word length economy isn't really the first priority. They shouldn't get longer, but it's fine if they don't get shorter. There shouldn't be any combo letters like this, instead it should simplify reading and transcription. "Spelling bee" in Vaultian would be a pointless contest where anyone [with the intellectual capacity to pass the citizenship test] would need no effort to get a perfect score.
No ambiguity in spelling, although the 16-letter alphabet might have some in pronunciation.

Phonetics remain primarily similar to English. Although pronounced more British-style, closer to direct phonemic spelling like in Latin or Greek, and possibly slightly simplified to facilitate it.
It's not exactly a distinct language, but not just a dialect either. Maybe like if it branched off at Early Modern English, then experienced a lot of language-mixing, and its formalization started in an environment where business machines already existed.

The numbers 16 and (31+space) are not incidental, they are powers of two, as the alphabets would settle on what can be most efficiently represented in binary. Most working-class Vaultians can't even write, though they can type (due to typewriters, not phones). Typewriters usually print 0-6 dots representing the symbol in binary, not a pretty letter, as paper has historically been in deficit.

It's mostly recognizable as a derivative of English, still cross-readable. Up to 2-3 letters can be replaced (with the ones added in the formal 31-letter alphabet), no more. However, it has to be cut to just 16 letters, not any arbitrary number, and that leaves a choice of which to keep.
And shouldn't be a paper-wasting jumble of letter combos from stenotype, the 16-letter alphabet has to be fully self-sufficient and used by itself for most informal writing. I'm not sure if it's easy or even possible, but would hope that it is.

If you're looking to simply contract the letters used into their most basic phonetic elements, that's easy.

Approach it as a list of "defined" phonemes in English (Wiki is an easy source), and determine your minimum phonetic level. From here, start matching existing graphical letters to phonemes. You will end up with lots of redundancy (Which is a good thing, but you are looking to shorten it).

In a matter of an hour you could easily cut a few letters out, reduced phonetic definitions to a much looser tolerance which does not cause misunderstanding. If you want to change that more, fuck with the syntax a bit and make yourself a distinct dialect, or even the morphology. It's endless.

It's actually quite easy.

Creating an alphabet from scratch would be easier.

But working on English, the most common sound in English speech is the schwa (Say, basic "e"), and many vowels can be transitioned into schwa without causing issues, and are therefore not necessary as defined vowels. Right there, so much can be retracted to just "e" with slight variations of pronunciation being on a per-word basis, or not at all. Then take account of glottal stops (Hilariously common in English, often represented by " ' ") and bam, you've got "e" and " ' " to replace a lot of vowels, and letter combinations.

Edit: Or you could just reduce phonetic values to binary and be done with it. Why even have the alphabet in there at that point?

I tried and I can't get lower than 18-20 letters. Most are hard to throw out. Y,Z go on the basis of phonemic redundancy, then G,Q due to unacceptably ambiguous graphemes. W is just VV.
C and X are too basic to go, maybe S and K could with C=K, X=S (Greek spelling). But then X is taking on Z,X,S,C(s), too much for one letter. P-B is a weak candidate for reduction, but R has to remain, can't see spelling a lot of words without it.

It would be. But it's not historically plausible, since it had to evolve semi-naturally from existing or former languages, primarily early modern English.

OK, that might work. And I do need a hybrid letter-separator that has both a phonetic meaning and serves as a space/punctuation mark.
But which vowels exactly am I obsoleting here? A,E,I,O,U seem pretty distinct.

Binary is just a representation. It still represents an alphabet, even if it's written with dots. So the alphabet still has to be reduced to 16 letters to store it in 4 bits. And in this case it also retains an alphabetical representation.

Binary is just a representation. It still represents an alphabet, even if it's written with dots. So the alphabet still has to be reduced to 16 letters to store it in 4 bits. And in this case it also retains an alphabetical representation.

I'll go over the rest later, I have to run.
But a representation is an alphabet. Therefore, binary can be itself an alphabet. The only detail you need to know is how it's represented. Which, you already have (Dots), therefore, you have a unique alphabet.

Yeah, I get that. But that's graphemes, they are not the difficult part. What I mean is that I don't know what phonemes to associate with them. For keeping the language close to English, talking here in terms of English letters to assign the phonemes to, rather than binary codes.


I tried ABCDEFHIJLMNOPRTUVX (that's 19), that gives "cvvic braun focx jampx over the laxi doc". Thinking of removing "j" and "u", but they need replacements.
You've mentioned the idea of using a modifier, and I like it, it will take out one more letter though. But I'm not sure whether to have it double as an often silent letter ("h"?), and how specifically to use it, i.e. what it modifies. It's also preferable if that modifier can double as a word separator for readability (since there is no "free space" when stored in binary), i.e. if adding it at every word's end still lets it be useful.

Do you know, or are familiar with IPA?
If so (Or feel like learning), it should help. Just using an IPA chart for English (Whatever dialect you pick) should allow you to see what sounds are similar, and start cutting.

For example, I would cut "c" but keep "k" and "s". Then perhaps used "ss" as a modifier for the difference between "s" and "sh". So sit would be "sit" and shot would be "ssit" all other letters being the same. You can do similar things elsewhere. "kk" for "g", and so on. It keeps words normal-length, but cuts letters out.

Somewhat. Not completely. But it doesn't do enough help, because every grapheme in English corresponds to multiple phonemes.

That's pretty ugly, and definitely increases word length.
It is pretty obvious, but it's still a code, something I want to keep in check.
So, you don't think cutting the alphabet down to 16 letters without resorting to codes and without heavy ambiguity is possible?