sorry about the delayed response; I've mostly been to sleepy to 
concentrate ....

On Thu, 13 Oct 2011 12:17:05 +0100, R A Brown 
<[log in to unmask]> wrote:

>On 12/10/2011 23:33, neo gu wrote:
>> Thanks, at least I know that (1) my post didn't just
>> disappear and (2) it wasn't clear
>> The basic problem is that I can't decide on what subsets
>> of 2-syllable sequences (or single syllables for that
>> matter) should be allowed in roots. For example, should
>> roots be limited to (C)VCVC?
>> Does that help?
>Yes, it does.
>Personally I don't see what single syllables aren't enough -
>but I would not restrict (C)V(C) to C being a single
>consonant.  I would allow consonant clusters; but the
>clusters one might have as onset will not be the same as
>those allowed as codas.
>More details will depend on the structure of your
>protolanguage; if, e.g. the language is has fusional
>inflexions (like PIE) or an agglutinative structure, roots
>will appear with endings, so roots might have the structure
>(C)VCC- where the second C is the onset of another syllable
>which forms an ending/suffix.

I'd like the language after sound change to be somewhat fusional, but 
the protolanguage could be agglutinative. (C)VCC- is good -- I've used 
it elsewhere -- but didn't want to use it for this language, just (C)V(C).

>>> In your reply on the 9th you wrote:
>>>"original vowels: i u  @  a
> >> original consonants: p  t  k  b  d  g  f  s
>>> m n  l  R (both onset and coda)"
>As onsets I would certainly suggest /p k b g f/ + /l R/ -
>whether you restricted /t d/ to just combos with following
>/R/ or whether you allow /tl/ and /dl/ is up to you.
>Then there are all the possibilities with /s/ + C.
>Altho some languages allow prenasalized plosives as syllabic
>onsets, e.g. /ntu/, they are far more common as codas. Also
>/l/ or /R/ followed by another consonant are common codas.
>But really, in the end, the choice is yours.

all good suggestions for another project

>On 12/10/2011 23:54, neo gu wrote:
>> On Wed, 12 Oct 2011 14:04:37 -0400, Anthony Miles wrote:
>>> 4. I find it hard to believe that R (by which I assume
>>>  you mean the uvular "French r") has no effect on
>>> surrounding phonemes.
>> It probably does, but I haven't decided how much is just
>> phonetic.
> If you haven't decided, it would surely be better just to
> write /r/. Using /R/ definitely conveys the impression that
> you mean a uvular sound here, either a trill or the Parisian
> uvular approximant.

What I meant was I haven't decided if, for example e > E / __ R, the 
resulting sound is a different phoneme or just an allophone.

> Rhotic sounds, however, of whatever kind are notorious at
> affecting surrounding phonemes.


>On 13/10/2011 01:29, Roger Mills wrote:
>> Seems to me that possible syllable structure is entirely
>> up to you; make it as free or restricted as you like.
>> Sound changes of course wrought havoc on that system;
>> you may have read the final result on my website.
>   :)
>That's what happens in natlangs as well.
>IMO neo gu should concentrate first on the protolanguage and
>not worry about subsequent sound changes.  That's what
>happens in the 'real world'.  After all, our PIE-speaking
>ancestors did not say to themselves "What sound changes will
>our descendants make? We'd better pick phonemes accordingly"
>I know PIE has been *reconstructed* by working back though
>sound changes; but that's because no one wrote down PIE at
>the time.  Also, the reconstruction of PIE is possible *only
>because there are several quite different daughter
>languages.*  PIE is a reconstruction of something that can
>reasonably explain the relationship between the differing
>daughter languages.
>OK - neo gu could, of course, construct two or three (or
>more) conlangs and then try to derive a protolanguage from
>them.  That approach has been suggested before, but it is
>fraught with difficulties.
>IMO it is best to concentrate first on the protolanguage.
>Having got that sorted out, one can derived daughter
>languages; and the exercise seems to me only worth the
>effort if one's going to derives at least two interestingly
>different daughter languages.
>> Is stress going to be important ?
>Stress and/or tone must surely be important when deriving
>daughter language(s).  but, as i say, my advice is to
>concentrate first on the protolanguage.

My only comment here is to say the protolanguage exists only to 
provide depth to the single daughter language.


>Nid rhy hen neb i ddysgu.
>There's none too old to learn.