Engelang phonology chat - engelang messages

On Fri, Aug 24, 2012 at 12:03 PM, And Rosta <and.rosta@hidden.email> wrote:

Mike S., On 24/08/2012 03:13:

> Let compounds be merely a concatenation of the stems. It doesn't matter if, say, CCCC is ambiguous between CC+CC and CCCC, or CCCCC between CC+CCC and CCC+CC: compounding would be a purely mnemonic way of forming novel predicates, derivationally translucent.
>
>
> I know you disagree, but I don't think it harms anything to have a
> compound hyphen that derives non-fully-compositional compounds. There
> could also be other hyphens to produce the fully compositional
> meanings based on the suffix. Given Xorban's syntax with overt
> variables and binary operators, we could potentially avoid a lot of
> verbosity via such shortcuts. Something to put on the back burner for
> now, of course.

I'm all for something that creates brevity. But a compound hyphen reduces brevity.

There are going to be vowels in these consonant strings. Some of them can be hyphens. If you compound br- and nt- without a hyphen, you can get things like br@nt or b@rn@t or b@r@n@t. If you compound them with an o-hyphen then you get bront or b@ron@t. If you combine them in the opposite order you get non-hyphened n@t@b@r or n@tbr- and hyphened n@tobr-. If planned well, hyphens will not reduce much brevity in the CCC* system.

> Discussion of phonology is really a separate discussion, but since you brought it up, I'd go for:
> vowels [a e i o u y @], <a e i o u y> with /@/ unwritten
>
> The one thing that Zamenhof is universally acknowledged to have
> gotten right is the vowel system :) I am surprised that a native
> Spanish speaker has sanctioned this six-vowel inventory.

I'm not a native Spanish speaker, I'm a native English speaker.

I'm sure there's nothing that Zamenhof is universally acknowledged to have got right.

I believe there is wide consensus in IAL-land that /a e i o u/ is an optimum inventory for vowels (universal acknowledgement is a gross exaggeration of course). A lot of people mix /@/ up with other things, and this will probably be one of the most common segments in Xorban. It's fine by me, though. I speak English too.

> consonants, probably one for each remaining letter of the 26-letter alphabet, /b c d f g h j k l m n p q r s t v w x z/. Assuming Lojban phonetic values, that leaves values needed for <h q w>. Obvious phonetic values are gap-filling velar nasal and voiced velar fricative, and glottal stop, which Lojban has (but with very limited phonotactic distribution). for slightly better grapheme assignment, use <h> for [x], <x> for [S], <q> for [?], and <w> for [G] and <c> for [N] or vice versa.
>
> That's efficient, but ugly. Is it forbidden for loglangs to use
> Unicode? If I were competing for Prettiest Romanization Award and not
> worried about convenience, I'd make Lojban look like the eastern
> European language that it sounds like:

The attraction of choosing the graphemes first is that it's (arguably) easier to reach a consensus on having 26 graphemes (because that's what most versions of roman provide) than to reach a consensus on phonemes. Given the paramount goal of brevity without information loss, there'd be pressure to have the phoneme inventory as large as possible, where the limits of possibility are when the phonetic distinctions are too difficult for most people to maintain. I think it'd be hard to reach consensus on where that limit lies.

FWIW, Livagian has 21 unaccented consonants, 4 accented consonants, 7 unaccented vowels and 8 accented vowels -- 21 consonants and 7 vowels if you take only the unaccented. So that gives a sense of where I strike the balance.