If you look at actual human speech spectograms, and the like, it is virtually IMPOSSIBLE to tell where one word ends and another begins in normal human speech. That's why speech recognition software has beeen so traditionally tough to do. Whereas it's pretty trivial for the human brain, (since we work with several layers of processing at once, taking into account a lot of factors, such as anticipating what the likely next word is and so forth), it's much tougher for a computer. So, unless we want..to.. be... stukc...talking..like..this, we must make sure that spoken lojban works even for people who, like me, a son of the northeastern US, used to rapid speech production/reception can be understood unambiguously in lojban. Hence the emphasis on stress (or the stress on emphasis) and C/V rules for breaking up word boundaries.
--Mike T.
On Wed, Sep 7, 2011 at 2:30 AM,
<rdentato@gmail.com> wrote:
On , Ross Ogilvie <
oges007@gmail.com> wrote:
> The issue is that there is no whitespace in speech.
I would read {pa plise} and {paplise} differently putting a short interruption between {pa} and {plise} in the first case.
The fact that they are the same just put more strain on the listener that has to break up words in his mind without the help of that short interruption.
After all, we do have {.} so if I really meant two words, why didn't I say {pa.plise}?
But, I'm not complaining, just wanted to understand the benefits. So far it seems to me that this would help the speaker making him free to pronounce words as a single flow with a price for the listener that has to break things up without counting on the short pauses that the whitespaces imply (at least to me).