buffer vowels

[And Rosta <cbmvax!uunet!pucc.princeton.edu!ucleaar>]

For those whose gobs rebel at articulating _zbasu_ or _sumti_ or other such
formidable consonant clusters, Lojban allows one to insert any vowel that
is not part of the standard /e i u o a @/ inventory. John Cowan would use
[I].

Now I believe this is not feasible. I am very likely to hear John's [I]
as lojban /i/. Similarly cardinal 13 vowel like in English English _hot_
(written with inverted italic/chancery lowercase a) would get heard as
/a/ or /o/. Mid front round '[o"]' would probably be heard as (lojban)
/y/. The only vowel I'd be confident of knowing the speaker not to intend
to represent some standard lojban vowel is IPA [y], often written u" -
a rounded [i] as in french _tu_. But (a) this vowel has not been declared
obligatory for this function, (b) the vowel is not among the easiest to
make, and (b) it is hardly the sort of unmarked vowel suitable for
unobtrusive separation of incompatible consonants.

I wanted to don body armour and suggest that lojban syllable structure
be declared to be CV(V) (extra V for diphthongs only), with C slots
allowed to be empty word initially, and with V slots either filled by
/a e i o u/ or by - at the speaker's discretion - [@]. This gives a
much more elegant & practical phonology. BUT YES I REALIZE that it has
very severe ramifications in lujvo morphology (I don't know about
le'avla) but in my blissful and bumptious ignorance I nevertheless think
it is worth considering. From reading level 1 materials, the only bad
problem I found was in lujvo that would fail the tosmabru test (the
only way to legitimize *tosmabru is to make it _tosymabru_ - but on
my suggestion, the form would anyway be _tos(y)mab(y)ru_ and therefore not
remediable by y-insertion). An extra device would have to be introduced
to guarrantee that something is a lujvo. I think ttosmabru (= tytosymabyru)
would be unambiguous (where the first consonant is duplicated).

I should reiterate that I can't decide whether it is appropriate to
criticize apparent design flaws in Lojban. On the one hand one shouldn't
assume that flawlessness is intended, but on the other hand Lojbab and
Athelstan have offered this sort of criticism to Loglan, and Lojbanists
seem fussy about meeting certain design criteria like making ambiguity
avoidable, & creating a robust signal, etc.

No offence intended, & I stand prepared to be asked not to post this
sort of carp to the list.

-------
And