Rex May wrote:

> Let's see, suppose you have five vowels and
> three diphthongs (Call them all 'V').  And
> twenty consonants.  You then have 8 morphemes
> of the shape V, 160 of the shape CV,  And if
> you can end with a consonant, you have3200
> CVS's.

If every word begins and ends with a consonant,
these words will produce consonant clusters
that are nearly impossible to pronounce.

Larry Sulky wrote:

>So Txeqli's scheme is actually quite elegant...

Some Txeqli/Tceqli phonemes are difficult
to pronounce:

R as in RoaR
bx - bilabial fricative as in Spanish 
dx - as in English THen 
gx - fricative as in Urdu baGH 
kx - unvoiced version of GH 
tx - as in English THin 

Many Txeqli/Tceqli roots are too long
for compound words.

> 1) Begin and end all morphemes with vowels.
> ilomi does this almost completely. It requires
> the use of a glottal stop between like vowels
> at morpheme boundaries, but that is a natural
> inclination anyway. The overall word form seems
> plausible until a learner realises that it's
> not just some words that have this form, it's
> bloody well all of them.

This is the best way. Stressed antepenultima
helps too.

> 2) Begin _or_ end all morphemes with diphthongs, which do not occur
> elsewhere.

Interesting, but not compatible with
compound words because you would have 
to use the diphthongs as part of speech
markers and diphthongs are too long for
that. Efficient, terse language uses
part of speech markers that are as short
as one phoneme.

>3) Use heavy stress on the first _or_ last syllable of all morphemes.
>This requires that speakers be accustomed to the use of stress. Also,
>it's difficult to do on one-syllable morphemes. A variation is to have
>one-syllable morphemes use diphthongs, which do not occur elsewhere,
>for their vowels. Another variation is to eschew one-syllable
>morphemes (which would make you, Rex, very sad).

Stress alone does not suffice.

>4) Designate a vowel as an 'extender' vowel, used only in unaccented
>syllables and never elsewhere. Konya did this, in both versions. A
>word couldn't end until an accented vowel was encountered (not
>literally accented, but simply any vowel that wasn't the extender
>vowel). Rick Morneau mentions this possibility in one of his shorter
>monographs. Actually, he probably covers several of these options

Vowels are too precious to waste even one of them.

>7) Tone profile. For example, one-syllable morphemes have a level
>profile, multisyllables go high-to-low in the first syllable, maintain
>low in the middle syllables, and climb back up low-to-high in the
>final syllable. Subtle, but perhaps sufficient. Quite plausible.

Too difficult to learn.

>8) Fixed-length morphemes. Always one or always two syllables, say.
>Difficult to adapt foreign terms and can sound monotonous, but there
>are natlangs that are heavily monosyllabic or heavily bisyllabic. Tunu
>uses this approach.

It would be difficult to distinguish the
one-syllable words from the two-syllable words.

>One further complication: SSM versus SSW (self-segregating words).
>Most of these schemes can alternatively be applied to whole words
>instead of morphemes, but they can't handle both simultaneously, so if
>both forms of segregation are desired, it's necessary to have a
>compounding morpheme, or a compound prevention morpheme, or to use two
>of these schemes together.

The self-segregating roots/morphemes are
as important as the self-segregating words.
Both Ygyde roots/morphemes and Ygyde
words have self-segregating morphology.