Are you fond of the highly inflectional Old English, with its loose word order?
Does an agglutinative tongue like Finnish please you aesthetically?
Do you like Latin overall, but just can’t adjust to the lack of articles?
Well, this is the thread for you, then. Tell us what your ideal language would be like. I promise not to mock people who praise Esperanto.
It’s a trickier question than it might seem; different attributes in a language would be ideal for different aspects of language use.
For example, for the sake of poetry, I favor a loose word order, with the function of each word established by the word itself, not by positional grammar. Similarly, an extensive vocabulary with significant overlap in definitions is useful in allowing for variety of sound, meter, and subtle connotation. These traits provide for greater flexibility and expressiveness within a poetic structure.
For technical use, it would likely be better to have a highly rationalized language, with straightforward rules, clear and distinct definitions, and minimal exceptions.
Here’s an attempt at some rules that would accommodate both:
Position/word order is largely a matter of custom or convenience.
Words that can fulfill different roles in a sentence will have their function indicated by prefixes or articles. (Alternatively, infixes are an option, but I feel they would unnecessarily complicate learning the language.)
The vocabulary will be extensive, including a substantial selection of near synonyms for terms related to complex subjects.
A simple idiomatic indicator, perhaps a leading interjection, will be available to indicate that one is using terms in a strict and literal sense, stripped of connotation.
Orthographic exceptions will be minimal, ideally limited to loan words from other languages. If the loan words come into common usage, their spelling will be normalized and they will be subject to the usual prefix/article rules.
The grammar will be genderless.
Pronouns will include both gendered and gender-neutral pronouns. The gender-neutral pronouns will be applicable to people as well as things.
There is always a catch: No matter how perfect a language is, the speakers will gradually evolve it, vowel shifts and so forth, so that the consistencies trickle out of it, and you wind up with exceptions to the rules. Like Old
English, where -ed past tenses were formed as separate syllables, and people just got tired of saying them, but in some instances it was easy to liase the -ed, and in some cases it was not, depending on the adjacent consonants, and some had to be phonetically converted to /t/, and now there are “rules” to be learned, and people asking how to form a perfect language.
Mandatory evidentials.
Don’t you think AM radio talk show hosts would just use the evidential signifying that they know it for a fact and can prove it, and life would go right on as it does without the spurious evidential? The presence of evidential implies an honor system as a societal prerequisite…
Well, if it were dropped into the language by the language fairy, sure, but the OP wasn’t asking how to fix english. My ideal language has mandatory evidentials and would require speakers to mark the level of confidence they have in the verity of their statements.
Also, we-exclusivity. I want a we that lets me say to someone “We’re going to get ice cream” and have them know for a fact, I am not including them in the ice cream group.
I’m not extremely interested in language, but I have learned two foreign languages (though learning the second one forced almost everything I knew about the first out of my head).
I’d definitely choose a language with little-to-no difference between the written word and the phonetic pronunciation, though I can accept the written characters having a little bit of leeway, to allow for dialects, without having to have dialect-specific dictionaries. But Japanese and English are both probably the worst languages to learn, simply because they’re almost impossible to read. No helpful rule for reading is effective for a useful percentage of cases, and you need to be able to read to truly advance in the language.
I would definitely shoot down any arbitrary distinctions, like whether my toaster is male or female. That just doesn’t add anything of value.
I would like to allow greater flexibility of abstract vs. precise. In English, there’s the distinction of “a” and “the”. Japanese has no such thing. In English, you’re forced to decide to be precise about whether it’s “a pen” or “the pen”, even though, from context it’s often going to be clear without the verbal declaration. Only when you need it, in Japanese, would you add the words “rei no” to indicate that there’s a special pen that overrides all others. That seems more reasonable to me, once you see how rarely you actually need that sort of specificity.
I also like, in Japanese, that you can drop any part of the sentence that we’re assumed to know. If I’m going to go walk the dog, in English I would need to say, “I’m going to walk my dog.” In Japanese, you can crop it down to, “Dog’s walk”, and it’s presumed that your full sentence was, “I’m going to perform the dog’s walk”. Japanese is sufficiently overly-verbose that they have to drop chunks of the sentence, or they’d never finish saying anything, but it does point to the ability for one to strip a sentence down to just the parts that would be unclear from context.
ETA: I’m curious to try learning Lojban
I was reading something the other day and came across a language (can’t remember which, sorry) where the usage specified the speaker’s knowledge source:
“I saw it with my own eyes”
“I saw evidence of it with my own eyes”
“Someone I trust saw it with their own eyes”
“I got no clue, so this is iffy - buyer beware”
I thought that was quite nice.
You could do that by dropping the “we’re” and saying “Well, going to get ice cream.” The toodle-oo is implied.
But is just the speaker going, or the speaker and his cohorts? Unclear.
The way mechanisms of poetry differ between languages must be a fascinating topic. Does anyone know a good webpage discussing it? English builds metric forms from differing stress (tone) while Latin uses syllable duration. Thai, with multiple long-duration tones, has more intricate forms.
If I were to create a fictional language, I think anapestic or dactylic meter would be obligatory in common speech, but judicial pronouncements, and even witness testimony, would always be given in iambic!
The main problem with English is that its orthography is not just very old, but specifically predated the Great Vowel Shift, which screwed everything up. Aside from that, Japanese and English have the opposite problem. Japanese kanji generally have several readings, corresponding to separate borrowings from Chinese. (And sometimes they just have different readings because nothing in the Japanese language is easy.) Turning a kanji into pronunciation is the tricky bit; on the other hand, there’s at least an order of magnitude more characters in Chinese than Japanese. Reading a written English word is nontrivial but not especially problematic; on the other hand, spelling bees exist.
The best explanation I’ve seen of linguistic gender is that provides an extra checksum for a sentence, or at least that it’s the vestige (at least in Indo-European languages) of another distinction (in PIE, animate/inanimate) that at least had some semantic or morphological (in PIE, an ergative/absolutive setting) content. But I agree: I would be totally happy to ditch gender and similar nonsense. Irregular verbs are inevitably going to pop up from sound changes (as well as processes like suppletion), and shifts in pronunciation are going to wreck your carefully-planned orthography, but there is just no real benefit to having gender in language. Not in our carefully designed language, at least.
Also, as a beginning Japanese student, I’d love to get rid of counters.
It’s not as general you describe, but I think pro-drop languages are generally the ones where verbs either differ for every subject (e.g, Latin) or are the same for every subject (e.g., Japanese). Languages that are somewhere in between tend not to be pro-drop. There are exceptions, but it’s a pretty strong correlation.
Eww. 
Lojban misses the entire point of language.
I’ve seen these pop up in a lot of conlangs, and the people who like them tend to really like them. What’s their benefit?
By the way, here’s a great site for outline the basic sounds, grammar, etc. of an invented language (and making it come out as a genuinely different language rather than a retread of English with weird spelling). I’d also recommend In the Land of Invented Languages, which is an fascinating book about how and why constructed languages are made.
(Also, sorry for the multiple posts.)
I don’t think you can get around the fact that, to some extent, language is a reflection of culture, and if you imposed a language on a culture that had not evolved the language to suit its own needs, you would either force the culture to change, of leave your language open to modifications that the culture would impose, deviating from the “ideal” supposed by some external language inventor… In other words, how many words for “snow”?
The International Scientific Vocabulary is evidence of this, in that every language is forced to adapt itself to constantly morphing needs that come with societal change. An example that comes to mind would be the introduction of electricity, which made each language decide whether one “opens” a light, or “turns on” a light, or some other construct based on forming a new analogy based on pre-existing language patterns.
The inventors of the perfect language would face problems like the naming of colors. How would the spectrum be divided into color groups? Would blue and green the the same word? Would dark blue and light blue be the same word? Would you create a few color names and use attributives, or would you have lots of free-standing color words? Would you use the same word for a color and a fruit (orange), or have different words? In fact, how would the ideal language deal with homonyms at all? Or legislate the exact boundaries of what a word denotes.
Well, back to the old drawing board.
You are exactly correct about the intent of the OP.
I’m in favor of clusivity, though not as the default. I’d like the default first-plurals to be akin to English we and us, with suffixes to mark inclusivity and exclusivity.
Take your diachronic blather back to Cardiff where it belongs, Welshman.
I don’t consider myself a linguist (language scholar), though I did earn and use a certificate to teach English as a Second Language, nor a linguist (fluent speaker of multiple languages), though I studied Spanish, Russian, Japanese and dabbled in speaking Farsi, French, and Chinese. So I don’t know if I’m really qualified to answer the question. Nevertheless, an ideal (to me) language would…
Have no irregular verbs. But verbs (in particular for this point) and conjugations thereof tend to fade as they fall out of usage. Thus, a verb like BEATIFY [to declare holy; a religious act, notably among Catholic Christians, which is one of the steps in determining a particular person is a saint] is conjugated (present/past/perfect) beatify/beatified/has beatified while, in contrast, a verb like GO (which is used more frequently among more general populations) is conjugated go/went/has gone. Someone once told me (as part of a Trivial Pursuit answer) that Yiddish, the official language of Israel, was a prescribed language with no irregular verbs. However, the same person also told me that English and Hebrew are more commonly spoken instead (worldwide and in Israel), so perhaps the lack of irregular verbs didn’t do much to help it gain popularity. [That was 30 years ago; maybe Yiddish has become more popular since then?]
Have a full set of distinct vowel characters. Perhaps adopting the full set of International Phonetics Association symbols would help. I realize, however, that doing so would replace the 5 familiar vowel characters in the English language with a set of 36 symbols, not all of which would be actually utilized by real speakers because real people are basically lazy (even when it comes to uttering noises).
Have a full set of distinct consonant characters. Again, adopting IPA symbols would help, or mixing symbols from Unicode or the Greek/Cyrillic as well as other ligature and rune collections would be useful. Again, this would end up replacing a lot of familiar symbols with a lot more symbols that are both familiar and unfamiliar.
There would be benefits of 2 & 3 to off-set the problem of having to learn a shitload more symbols
A) One symbol = one sound ONLY, ALWAYS*. Not only does this obviate the need for symbolic clusters (ch in character) that have different pronunciations anyway (ch in chalice; ch in chartreuse); but it also reduces overused symbols (hard c in control; sibilant c in cervix) and allows for redeployment of some odd symbols that get seem to be employed because it was too much effort to find the more suitable symbol (G in general, gray, and genre).
B) Silent letters (p in pneumonia, e in maze, u in plague) would no longer be necessary.
C) Orthographic exceptions would go away – I before E except (immediately) after C or when used in…
D) In fact I think that, once the whole alphabet was known, spelling errors would go away – or at least be extremely reduced – because a careful, correct, and accurate pronunciation of a word would allow ‘reverse engineering’ of the spelling.
E) CaPiTAliZATion could be ignored. Most English-speakers have roughly 36 consonant symbols memorized as consonant pairs (uppercase vs lowercase – G|g M|m etc.) and then a few that look about the same in both cases (V|v S|s etc.). If one character = one sound, then there are more symbols that we already know that can be used for the phoneme-a-bet (which means we really don’t have to learn that many more symbols).
Be able to adopt and adapt terms from other cultures as well as science & technological progress and social evolution. [In other words, there would be no Language Purity Review Board prescribing or prohibiting new terms in the lexicon.] However, since the language’s symbolic representation is designed to accomodate all phonemes, terms brought in from other cultures would be spelled with the language’s symbols rather than the originating culture’s characters – the borrowed words would sound the same but could look extremely different. Naturally, terms would also fall to disuse and that would be acceptable as well.
Have no tonal/inflective distinctions between words. This is just because it’s what I’m accustomed to using - in contrast to Mandarin Chinese, for instance, in which a rising-toned “Ma” is a completely different noun than a flat-toned “Ma” - and tone shaping indicates other nuances like inquiry or sarcasm, or imperative etc.
Use punctuation marks with one mark = one meaning, always. Most importantly, that would mean possessives would be indicated with something different than contractions. Also, I think it is helpful for some punctuation marks (like question indicators and interjection indicators) to be placed both at the beginning and the end of sentences so that a reader is prepared to modulate his tone (see above) well in advance.
Be 99.9% gender-neutral. Unless specifically talking about anatomical matters, I tend to think the assumption of gender neutrality in a language would help to fizzle out that symbolic-interactionist bone of contention.
Have distinct breaks between words. This was a pain in the ass for me when I was learning to read Japanese again (in Japan) and I didn’t realize how much I really like the crutch (from my college textbooks) of seeing spaces between words to help me know what I’m reading. Furthermore I think sentence, paragraph, section, and chapter breaks will be required as well.
–G!
You can see I come from a publishing/copy-editing background, since I’m focusing mostly on how the language reads.
*With a minimum of 36 vowels and more than 21 consonants, there’s going to be a ton of symbols to learn. I think the number of consonants can be cut by almost half simply via the use of a mark (maybe an umlaut or tilde or something like that) to indicate the presence of voice in a consonant, thereby easily distinguishing what we know as the pairs T|D, K|G, P|B, etc. Combine that with dropping the uppercase/lowercase distinction and I bet we already know all the symbols; we just have to associate some of them with new phonemes.
I’m no linguist, I just knows what I likes 
My ideal language would have no weird guttural sounds like the first syllable in “Haunnukah”. It wouldn’t have that weird pronunciation like the last syllable in “Martin”. There would be a phonetic spelling that would be enforced strictly, all words must sound like how they are spelled with no exceptions. It would not be like Chinese, where a word is essentially a picture and gives you no clue in how to pronounce it. It would flow well by useing lots of l, n, s, and m sounds. There would be no exception rules like how “ch” is pronounced differently than the letters that make it up suggest. Such sounds would be made by its own letter. It would be descriptive, where pre and suffixes have meaning that can be derived if you know the rules, and NO stupid male and female type words like “el” or “la”. Those are good rules to start with