Hallo conlangers!

On Saturday 04 August 2012 09:20:47 R A Brown wrote:

> On 03/08/2012 23:56, neo gu wrote:
> > I've been trying to make C1 a non-verbose language.
> 
> Presumably to make a language at least as concise as Lin.
> 
> > I've been trying to come up  with ideas for doing so in
> > general. Here are some, not all of which have been
> > applied to C1:
> > 
> > • use a large # phonemes
> 
> Not a new idea, of course. Lin has a large phoneme
> inventory, as does Ithkuil and some others.

R. A. Heinlein had the idea in 1949 when he wrote his story
_Gulf_ featuring a language named "Speedtalk", though as far
as I know, he never actually worked it out.  The idea was to
use a huge phoneme inventory in order to reduce all morphemes
to a single segment in length.  I haven't yet seen a working
language that implements this principle; but I am going to
try it out in Quetch.  Using a lexicon based on that of Toki
Pona, I expect to need about 120 consonants for the content
word roots and 20 vowels for the relationals.  (With such a
limited set of roots, the language will have to take much
recourse to compounds, but if each member of a compound is
only one segment long, even complex compounds will not be
all that long!)  The "purpose" of Quetch is simply to try
out how well the speedtalk idea will work out in practice,
and to measure how much brevity is actually won over natlangs
that way.

>       There are also
> some natlangs with large phoneme inventories.

There are.  The 120 consonants and 20 vowels I am expecting
to need for Quetch are many, but not entirely beyond the
range found in natlangs!

> > • don't write @'s and ?'s
> 
> Presumably you mean schwa and the glottal stop.  If _two_ of
> your phonemes aren't going to be written, then methinks
> you're going to cause confusion.  How will you know when
> these phonemes should occur and whether the unwritten
> phoneme is schwa or a glottal stop?  Why have them if they
> are not written? or do you mean they will be marked by
> 'white space'? (White space is pronounced in Lin)

Apparently, he will use some sort of phonotactic rules (see
his reply).  In Quetch, a schwa can be inserted to break up
unwieldy consonant clusters.  The glottal stop, on the other
hand, is an "escape symbol" for proper names: the name is
encapsulated in two glottal stops which do not occur elsewhere
in the language and are also not allowed in names.  The name
thus encapsulated is grammatically treated as a content lexeme.

> > • use zero for the most frequently used markers
> 
> If more than one marker is marked by "zero", how will a
> reader know which marker is "understood"?
> 
> > • fusion
> 
> Yes - but can you be more specific?
> 
> > • use word order to reduce # inflections needed
> 
> Isolating languages cannot have any inflexions at all; they
> are not actually needed!

Indeed not!  The Chinese managed to build an impressive
civilization without this "mortar"!

>       But, unless you have some form of
> self-segregation of morphemes (and I think a very concise
> language will need this), you'll finish up with a lot of
> white spaces.

Quetch is trivially self-segregating.  Apart from encapsulated
proper names whose extent is explicitly marked with glottal
stops, all morphemes are exactly one segment long.
 
> > • reduce  redundency: don't mark person and number when
> > corresponding phrase is present?
> 
> I.e. don't mark things if they are obvious.  You can reduce
> redundancy to nil   ;)

I don't know whether one can *always* reduce redundancy to
nil.  A zero-redundancy language would be one where any string
of segments is a meaningful and grammatically correct utterance.
I don't really expect to reach that point in Quetch!
 
> > • Each lexical class specifies how the 0-marked and other
> > inflections are interpreted.
> 
> I.e. like a loglang, each lexeme is a predicate and
> indicates the arity of its arguments.

Like my (dormant) X-1, where the arity of a predicate word is
indicated by its length (an N-ary word has N+2 segments).

>       Surely this alone
> means that inflexions are not needed.

Yes.

--
... brought to you by the Weeping Elf
http://www.joerg-rhiemeier.de/Conlang/index.html
"Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1