Re: [lojban] gizmu

["Lionel Vidal" <nessus@free.fr>]

> This is not an example of speaker laziness, but of bad language
> design; encoding meanings into sounds that are not well-suited to
> the organs that produce them.  A properly engineered language would
> not "suffer" from such assimilations, but would assume them as a
> requirement for its design.

>From a strict phonetic point of view, this seems unrealistic: all
known languages used in speaking meet, given time, some
phonetic erosion (assimilation is just an example of what can happen).
You could say that no language has been yet properly designed, but
even if could map all possible erosion mechanisms (something impossible
in practice), I am not sure that the phonetic space will be large
enough to design something that could not be eroded.
(and this language would be very difficult to articulate, with any
consecutive phoneme having an articulation scheme different enough
to prevent erosion:  fluency would probably requires surgery :-)

But to come back  to assimilation, this is not really a default as people
tend to not assimilate what they feel will result in a practical
morphological or semantical ambiguity. This is the main reason
for the assimilation rule exceptions.

-- Lionel