Theta wrote:JeremyHussell wrote:
Theta wrote:I don't think succinctness is really an important feature, just something that happens to be a characteristic of certain languages. Inuktitut tends to have very long words for common expressions and it does just fine.
Depends on your definition of 'important', I suppose. I'm interested in this feature, and ways to achieve it. That's pretty much the only justification a conlanger needs, right?
Well maybe 'necessary' would be a better word. I think that if someone finds their conlang to be not as succinct as some of the languages they know, it's not really a *problem*.
For most conlangers, it's true, succinctness is not important. But apparently it's important to the OP, so I guess we should just leave it at that.
I like my conlangs to be succinct too, as it happens. My main conlang Rammy is more about being clear and succinct in a technical way, whereas my secondary conlang The Choir Conlang is more about being poetically succinct - laconic, if you will.
cromulant wrote:JeremyHussell wrote:Has anyone ever read anything about correlations between language features and how widely spoken the languages are? I feel like I'm about to get into an argument based primarily on lack of evidence, which I'd like to avoid if possible.
...there are none.
I doubt very much that Mandarin has changed significantly just because it has a lot of speakers. English is a little special, because it is widely spread as a second language. There are some studies as to if and how this affects it. There are also studies of nearly extinct languages, to see if they are different somehow.
Let's see, on topic...
I think your list of strategies was quite good actually. Some of the strategies are not very good, but I don't think you meant to imply that.
I have been wondering about what happens when you have more allowed syllables. Clearly the number of syllables per second goes up, but does it go up enough to keep the data transfer rate constant? Or is there some number of syllables that maximises the rate? Should be easy enough to measure.
I think one thing that would help is to have a lot of features. Not a lot of degrees of features, because that makes you have to slow down for clarity, but a lot of separate features. One notable example is tone. Would you have to slow down if you add tone? Maybe, but you can think of it in another way: Suppose you have a non-tonal language, and then you add tone to each syllable. Surely that should increase the clarity? Which means you would be able to increase the speed.
Unfortunately I really don't like tone. Or stress. But in principle, I think they would help.
Fusion and irregularity can help make things more succinct. That might be an advantage of natlangs, at least compared to certain types of conlangs. Classical example: English "I am, you are, he is" versus Esperanto "mi estas, vi estas, li estas". The Esperanto version is easier to learn, because you can separate it into logical parts, so similar meanings have similar sounds. But English is much more succinct, not only because is has fewer syllables, but also because it has a bigger difference between them, so you can speak faster.
It's easy to do what Esperanto does, piling up suffixes for all your inflection needs. But fusion/irregularity is more succinct. In Rammy, I've used one strategy: Marking all inflections on little form words. That way they can be horribly fusional and irregular, and still reasonably possible to learn, since they are a limited number.
In the Choir Conlang, I've put all the inflections on the content words instead, particularly the verbs. They are inflected in such a way that each combination of root+form is not particularly similar to either other words with the same root, or other words with the same form - the very opposite of Esperanto or most other supposedly logical languages. In principle, it is more likely to be similar to a word which differs in both root and form. That minimises the risk of misunderstanding - since you are unlikely to misinterpret both root and form at the same time, you can talk faster. The problem is of course it's very difficult to learn.
Giving every sound combination a meaning is generally a bad idea, since you would have to speak extremely slowly. Redundancy (in a sense) is good. But you can still control your redundancy, if you're really serious about succinctness. You can, for example, make a set of five-phoneme roots such that no root has more than three phonemes in common. Or you can go to feature level and control things even more. You can even create the conlang equivalent of hashcodes. Horribly unnaturalistic, but perfectly doable.
One quite simple trick I think is "zero-syllable words". Try letting the word for simple things like "I" or "be" be just one sound, and not a vowel. Sure, it can be seen as affixes, but it might help to think of it as words.