Re: buffer vowel

[ucleaar <ucleaar@UCL.AC.UK>]

> >That's a reasonable justification for anything-goes. I think the reference
> >grammar presentation should make this much much clearer, though, since
> >it gives the misleading impression (a) that it is not the case that
> >anything goes, and (b) that lojban phonology is like natural lg phonology.
> And how is this unlike natlang phonology?  In real life, if I were to talk
> to you in English, and even in my facsimile of British English, I would
> almost certainly violate standards of phonology in a variety of ways.  This
> does NOT mean that "anything goes" - I cannot pronounce 'g' as 'd', 's' as 'l'
> 'r' as 'n' all at once without one of us getting pretty screwed up %^).  On
> the other hand, if I had a lisp, and substituted 'th' for 's', and also
> substituted all of the American English differences from RP, I suspect that
> you would still find me quite understandable.

The difference is that with natlangs you have clearly specified rules, etc
but that we cope pretty easily with the rules being broken, and rampant
rule breaking is constrained only by the need to remain comprehensible,
while with lojban there are no rules (of the relevant type) but the
need to be comprehensible means interlocutors must adopt ad hoc ones.

> In short, you can "get away with" regular shifts in the phonology and
> still be understandable.  We merely have defined that a certain range
> of shifts associated with the buffer, are not only understandable but
> by prescription "acceptable".

I don't see what you mean.

---
And