The Korean tried inventing an optimal alphabet a few centuries ago and it caught on. [0] Traditional Korean writing (in Chinese characters) has been almost abandoned.
The Korean Hangul alphabet doesn't include all possible sounds but I don't think that's really what you want. All those guttural consonants, tonal vowels, and clicks and snaps and buzzes that are possible add complexity, but the real breadth of words is mostly combinatorial rather than based on having a lot of letters.
The rule is: "Write as you speak and read as it is written."[0]
Regarding glyphs you have a number of problems. These include phonetic shifts. So for example, in English we have a word like courtyard. This is a compound word with two components, court (from French) and yard (from Old English). Both of these come from the same PIE root. Now this isn't too bad with court and yard, but let's look phonemes of another sort: what PIE kw- vs gw-. gw- became qu in Old English (hence words like queen), while kw (which becomes qu in the Romance languages) becomes wh (what, where, whorf, whirl). This is still arguably separate from a w sound as most people I know pronounce where slightly differently from wear.
Combine this with the fact that we have lost letters (thorn and eth, primarily) which refer to distinct phonemes which we now use th for.
To come up with an optimal language, I think you'd have to make it fully prescriptive and heavily spec'd out (like computer languages are) but this isn't the way human language works. I suppose someone, probably me, has to point this out, and if you didn't get the subtle jab at discussions of pronoun cases, please re-read this sentence, slowly.
There comes Chinese characters, which are exactly the opposite of your definition of "optimal": Chinese characters are not an alphabet, they do not carry phonetics, they carry meaning. And, guess what, this is arguably better.
Why? Because phonetics tend to shift quickly. In a few decennies any spoken language has shifted, vocalisation is different, especially when the language is used by loosely coulped communities. And then you have to adapt spelling to new pronunciations. But what do you do with old books? And how do you keep the language united? Chinese characters, by allowing different pronunciations without changing their written form, do fix this issue. This has helped China to stay centralized for a long time, and share one common continuous cultural ground.
So, if optimal means "can be shared by more people for a longer period of time" maybe the optimal language would need to add an articulation, and represent meanings instead of sounds.
But I would think (and remember reading somewhere) that phonetics shift more quickly where there is no written record of the pronunciation. Chinese has dozens of mutually unintelligible dialects (Mandarin, Wu, Cantonese, Hakka, etc.). Maybe this fragmentation would be much less if they had a phonetic writing system in the first place.
> In a few decennies any spoken language has shifted, vocalisation is different, especially when the language is used by loosely coulped communities. And then you have to adapt spelling to new pronunciations. But what do you do with old books? And how do you keep the language united?
English (or Spanish, etc) is a real-world example of a mostly phonetic writing system, and I don't think it is that bad. We can still read old texts, even though our pronunciation of the words has changed vastly. If you go back far enough (e.g. to the time of Chaucer in the 1300s), it becomes hard to understand what is being said, but practically speaking I don't see this as a big problem. In contrast, I think it's very important to have a writing system that is easily learned, so that it is accessible to the entire population (not just those able to invest lots of time & effort).
That's because at that time they still wrote phonetically. Today the spelling of some languages, like English especially, is fixed and immutable yet the pronunciation diverges.
Think about equATION and pronunciATION (or any other word ending in -ation that has "ʒ" instead of "ʃ").
There is also the reverse effect. Spelling influences pronunciation. Especially the sounds of single vowels (a,e,i,o,u), their sound being used frequently when spelling, becomes a guide to pronounce unfamiliar words, making words like "fungae" sounds nothing like what the latin word used to be (and how it was probably pronounced by english speaking educated people in the middle ages).
That said, even a crippled phonetic alphabet is certainly easier to learn than ideograms. However English is at the worst end of the spectrum. Japanese syllabic writing systems (hiragana and katakana) are much easier to write even if they have a few more glyphs. Unfortunately, even if you master -kana scripts, and thus able to write everything you want in Japanese, you'd be cut off from mainstream culture, including newspapers street signs etc
This is why Latin, which had such an alphabet, didn't split into mutually unintelligible dialects like Romanian, Spanish, and French, right?
> We can still read old texts, even though our pronunciation of the words has changed vastly. If you go back far enough (e.g. to the time of Chaucer in the 1300s), it becomes hard to understand what is being said, but practically speaking I don't see this as a big problem.
But this wouldn't be the case if we wrote English phonetically today.
For example, Sir Gawain and the Green Knight (http://quod.lib.umich.edu/c/cme/Gawain?rgn=main;view=fulltex...) is one of my favorite pieces of Middle English literature, and exactly contemporary to Chaucer. It's hard enough to read without some knowledge of Middle English but you'd be totally lost if we wrote nait instead of knight.
The different "dialects" are really different languages altogether (certainly as different as the various Romance languages, for example). It seems unlikely that a country as vast and diverse as China would have become monolingual just if they'd used a different writing system.
But the trick is that when written on paper these dialects are mostly mutally intelligible.
In the other hand, English and lots of other languages (including my native language Turkish) are synthetic languages, so they have affixes, and they do not adapt well to logographic writing systems (class of Chinese characters) in my opinion (I am not an expert on linguistics). My reasoning is not the cognitive overhead that's caused by memorizing tons of characters, but overhead of encoding/decoding the affixes into that type of writing system.
Japanese is an interesting example in this case, it is a synthetic language and has conjugation etc. so Japanese people use Kanji (Chinese characters as used by Japanese) to write word root and Hiragana to write conjugation, function words, spelling etc.
But that need to memorize IS a cognitive overhead that is hard to dismiss. The Japanese have to learn thousands of Kanji from their early years (of elementary school) till well into high-school! And I would say that the encoding/decoding is actually an overhead only until it gets automatized in subconscious mind.
In languages where letters are mapped to sounds, even after phonetics changed a bunch, the written and spoken forms are huge cues to each other. (There are exceptions).
If a new word came to be in English and it was written, for example, "scaramara", two fluent speakers who only saw it in writing would probably guess similar pronunciations for it, to the point that one would understand the other's speech. I suppose something like that could not happen in Chinese. Is this assumption correct?
So yes, learning Chinese or Kanji has a steeper learning curve, but no, it is not that hard, and in fact from context and shape you can guess a possible pronunciation and meaning for unknown characters, sometime.
I've always felt that I need to memorize both the written and spoken forms, when I study English.
Another interesting case is the Cherokee syllabary:
> Once Albert Gallatin saw a copy of Sequoyah's syllabary, he found the syllabary superior to the English alphabet. Even though the Cherokee student must learn 85 characters instead of 26, he can read immediately. The student could accomplish in a few weeks what students of English writing could learn in two years.
There are many conlangs for various purposes. Esperanto and Interlingua are useful for facilitating communication between people, acting as geography-blind APIs. Lojban is useful for communication between humans and computers, as well as fostering a strict logical mindset similar to the effects of working with programming languages, mathematics or symbolic logic. Klingon, Elvish and many other languages of various semantic/syntactic comprehensiveness are created for aesthetic and/or fictional purposes.
I've been working through a new conlang idea in my head the past few days. It builds off of Lojban in that its tenets are (1) universality, and (2) ease of I/O by both humans and machines. But it adds onto that (3) an easy-to-use, potentially unconscious system of real-time abstraction.
We do this to some extent with all human-human communication, when we inflect or affect or twitch or sigh or use other non-verbal linguistic cues to impart semantic value onto entire complex utterances. The way that a hyperlink imparts the contents of its destination onto the meaning of the information between its opening and closing tags.
This could be significantly sped up through use of more standardized multi-sensory linguistic components, either attached directly to individual people or attached as a metadata layer readable by nearby people through an AR technology of some sort (glasses are the obvious choice, but let's not limit ourselves).
For instance, a speaker could say the word "lu" multiple times in the same monologue. Those around the speaker would grok the intended meaning of "lu" each time due to whatever technological cues were broadcast along with it. The cues could be a string of words on a HUD or a color tinting the image of the speaker or a soft harmony in the listener's ear or a distinct pattern of buzzes from piezo elements in an ankle bracelet or even direct electrical stimulus causing a particular mental state.
I'm getting ahead of myself.
The point is that this is the sort of thing I imagine "language optimization" to be. And this example is just scratching the surface of potential optimizations, because any optimization that allows for general abstraction opens up all the doors of Turing completeness.
Esperanto and Interlingua are patches. They are certainly useful for some things[1], but we can't know what they're optimal for unless we're also experimenting with entirely new systems not based solely on natural languages. We need to be able to treat real-time language as scripting — anything less and we'll be limited by the language rather than by our cognition.
[1] Try reading this wikipedia page. Note how your reading speed increases exponentially as you go down the page: https://ia.wikipedia.org/wiki/Interlingua
Problem with that is that it leaves very little room for error detection in the language -- everything is a valid statement.
Personally? I think English is insanely stupid, inefficient, and dumb. I speak it because I have to, not because I want to, because everyone else does it. If you think trying to get breakout success in a market with, say, a software product is, try thinking of a way to create an "optimal" language (ie, glyphs only trace to possible human vocalization, but it covers the entire range - English doesn't, for example. And your words are a huffman tree - the most popular ones are the shortest, though that can only ever be realized when the language is conceived since the common vernacular shifts over time).
Now try to figure out how to get that market penetration.