Re: [engelang] phonology - engelang messages

On Mon, Sep 17, 2012 at 3:15 PM, And Rosta <and.rosta@hidden.email> wrote:

> Possibly. I'd be interested to see whatever you cook up.

One idea would be for the variable-marking Vs to be intercalated among the Cs, with /@/=o'e (& allowing /@/ to bear contrastive tone). This supposes that function words contain one C and things with multiple Cs are predicates. Orthographic spaces would distinguish /C@-lo/ <C > from /C@-hi/ < C>.

The virtue of that scheme is that it saves syllables by not requiring overt o'e that is present only for grammatical reasons and by replacing @s by Vs rather than adding Vs.

Interesting. So the number of consonants is at least as great as that of arguments, and usually greater since most predicates have one or two arguments. The consonants without arguments always have /(@)/. How would places be counted? They could start from the first consonant go forward or from the last going backward i.e. patf = A's a father, ptef = E has a father versus ptfa, ptef.

One possible problem is that /@-lo/ unlike /@-hi/ can never be elided -- maybe put X1 in the final position in order to try to forestall the temptation to do so (X1 the least likely argument to be elided)?

> But I do think we should be thinking about the issue of tone before investing too much effort in working on the current morphology that's based on toneless phonology.
>
> The morphology is so simple and the number of closed class morphemes
> so small that I don't think it would take any Herculean effort to
> give Xorban a new "skin". Morneau used to tinker day to day with
> parts of his relatively more complex MTIL, and then every few months,
> completely burn down his morphology and rebuild it from practically
> from scratch. The only thing that killed him is that each time that
> he did that, he had a substantial vocabulary that he had to refit
> into the new morphology.
>
> And that's where a real rub is: We would need to finalize the
> phonology and morphology if we were to start designing the lexicon.

We could design the lexicon, but leave the phonological shape of roots unspecified. That is, you'd decide on the meanings and on the adicity and possibly on whether the stem should be a compound of certain other stems, with the phonology left till later.

That's a good idea. It wouldn't hurt for people to start thinking about that now.

> > <x, q> for /G, ?/ is less weird than <'> for /h/. Also less
> > unwarranted.
> >
> > IMVHO they're all weird.
>
> <'> has the additional weirdness of not being a letter, tho ther's orthographic precedent of it representing a schwa interconsonantally and [?].
>
> Maybe it's worth considering <'> for the schwa and <q> for [?]. Then
> in my compounding scheme, the buffer vowel [y] would always be
> implicit (and often silent), and the hyphen vowel [9]/[@] would be
> represented by a lightweight punctuation mark.

I guess names would be delimited by <q> only then?

What would be the point of having any orthographic representation for schwa? It never has a contrastive function. Oh hang on, which version of which language areyou talking about? Which has [y] for buffer and [@] for hyphen?

Yes, sorry, I keep changing which scheme I am talking about.

> As far as <h>, there is some precedence for it standing for [h\], so
> maybe that should be the letter assigned to [G], if that phoneme
> where added, leaving <x> for [h] or [x].

<x> for [x, h]. Is [h\] a pharyngeal fricative? If so, that's hard to pronounce in a way integrated into the rest of the speech stream, and unlike [G] it doesn't fill a systemic gap.

[h\] is the breathy-voiced glottal fricative. Czech uses <h> for [h\] which contrasts with Czech <ch>=/x/=[x]. Czech ostensibly lacks a /G/, but one could arguably analyze Czech <h> as a lenited form of /G/ -> [h\] in which case Czech <ch>=/x/=[x] would have its expected voiced counterpart in <h>=/G/=[h\] after all.

What I think you are proposing is: the phoneme /x/ would be [x, h] and the phoneme /G/ would be simply [G]. Czech allows one to make the case that <h> has a natlang precedent for representing the phoneme /G/ whereas <x> seems to have none. As for <x>, we have seen that used for [x] in Spanish.

Given the above, <x> for /x/ = [x, h] and <h> for /G/ = [G] probably has a stronger case than the reverse. In addition, it would preserve the value of <x> inherited from Lojban; I know that doesn't carry a lot of weight with you of course.

> Right. We both know where the Lojban list is if we want to talk
> about Lojban's specification.
>
> ObXorban: How would that last sentence be translated in Xorban? Is
> there some odd illocutionary operator that we don't have yet
> involved?

I'd use the ordinary assertion marker, "ci" I think it is. Is that what you were asking, or were you asking what would be the xorban for the whole sentence?