Received-SPF: pass (google.com: domain of paskios@gmail.com designates 209.85.213.170 as permitted sender) client-ip=209.85.213.170;
MIME-Version: 1.0
Sender: lojban@googlegroups.com
In-Reply-To: <CAGO2uOPcMjNogyRO8Wi-v+swm60xLR8X21=aykzmcMOh51e0-g@mail.gmail.com>
References: <CAGO2uOPcMjNogyRO8Wi-v+swm60xLR8X21=aykzmcMOh51e0-g@mail.gmail.com>
Date: Wed, 20 Jun 2012 11:13:41 +0100
Message-ID: <CAPAJj9JOHh+4NxuL+uB-mtppWZ1gR8zKMHHFrj2aoAA7QzVsAA@mail.gmail.com>
Subject: Re: [lojban] Are Natlang the best case for entropy in communication ?
From: tijlan <jbotijlan@gmail.com>
To: lojban@googlegroups.com
Reply-To: lojban@googlegroups.com
Precedence: list
Mailing-list: list lojban@googlegroups.com; contact lojban+owners@googlegroups.com
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
X-Spam_score: -0.7
X-Spam_score_int: -6
X-Spam_bar: /

Below are my notes I've been putting down that I'm posting just now.
Don't bother to respond to where it overlaps with other jbopre's
points.


On 18 June 2012 10:52, Escape Landsome <escaaape@gmail.com> wrote:
> This is achieved by natlangs, by sparsing all the existing words (for ins=
tance, adverbs) in such a "morphologic space" that has very few collisions.

Also in a syntactic space. "To get hold of X", for example, can be
seen as a syntactically diffused equivalent of "to take X" (the latter
is sonically shorter and can therefore be harder to pick up in a bad
speech environment).


> But this is not the case of lojban words, for instance so'a, so'e,
> so'i, so'o etc. are very near of each other, and, assuming you don't
> hear well the last vowel, you could infer something very far from what
> was intended by the other speaker.
>
> So, is not that something that is annoying ?

My practice of Lojban is mostly textual, and the sonic disadvantage in
question (if any) may have been offset by the words' mnemonic
advantage for me (i.e. the set of {so'a, so'e, so'i, so'o, so'u} would
be easier to memorize than the set of corresponding natlang words with
incoherent dissimilar shapes).

The sememe-per-phoneme (information-per-sound) ratio, or information
density, can be as much relevant to the communicability of spoken
words. It would seem: the higher this ratio is, the more information
listeners are likely to miss out from a speech in a noisy place or on
a bad phone line. Lojban seems to have a high ratio, with its vocally
minimal lexicon. But that isn't to say natlang utterances are
altogether equally easier to scoop out of noise than Lojban's are.
Here's an interesting study:

  http://www.time.com/time/health/article/0,8599,2091477,00.html

  information density | average spoken syllables per second
  .94 | 5.18 -- Mandarin
  .91 | 6.19 -- English
  .63 | 7.82 -- Spanish
  .49 | 7.84 -- Japanese

A higher information density appears to be countered by more-stretched
syllabism (more diphthongs & consonants in the case of English). But
does that help in a noisy situation? As a non-native English speaker
living in UK, I often have a hard time figuring out what people are
saying on the phone, on top of the fact that they come off as speaking
rather fast overall in spite of the supposedly stretched syllables.
Vocal elements are largely idiomatized, sometimes even gestalt-ish,
requiring listeners to make up or convert in their mind the elided or
altered sounds:

  [ha=C9=AA=CA=94 p=C9=91=CB=90=CA=94] --> Hyde Park
  [b=C9=99i=CA=94] --> beat
  [b=C9=92v=C9=99] --> bother
  ...

This natlang phenomenon... Is it entropic? Is it convenient?


On 19 June 2012 09:02, Escape Landsome <escaaape@gmail.com> wrote:
> I understand that lojbanists use to protect themselves against any
> criticism, as in a besieged citadel, but the point still holds :  <<
> if very different concepts are given very near phonological forms, isn't =
this a bad move ?

{so'a, so'e...}, classified as PA4, conceptually or functionally
aren't so miscellaneous. The same for {ko'a, ko'e...}, {fa, fe...},
{la, le...}, {noi, poi...} etc.

A group of "very different concepts", for me, would be something like:

  many
  gentle
  monitor speaker
  week
  logical fallacy
  hat-trick
  central nervous system
  ...


> And : << Does not the fact natlangs do not have this problem generally
> speaking imply that they are more well designed than lojban on this
> particular point ? >>

I don't know that's a fact. What I know is that the following can
confuse foreign English learners:

  crush, crash, clash
  run, ran
  hit, hit
  decision, discission
  complement, compliment
  site, sight
  discrete, discreet
  flare, flair
  vice-president, a life of vice
  ...


On 19 June 2012 09:53, Escape Landsome <escaaape@gmail.com> wrote:
> Natlangs were designed, but the designer is a non-human (and
> non-divine) random process of natural selection.

Language use is largely filtered by human intelligence. Words to be
used (and recorded) can be selected non-randomly, even altered if
deemed necessary or effective for some purpose:

  shit --> shite
  fucking --> effing
  all correct --> ok
  laughing out loud --> lol

After all, modern English owes a lot to artistic writers such as
Shakespeare, for whom a linguistic expression could well be an
engineering process.

Many 'natlangs' have official regulators. French orthography, for
instance, has been occasionally reformed to improve both textual and
sonic consistencies according to the lexical families.


> Natural selection
> favorises random creations, but random creations naturally occupy the
> phonological space smoothly and in a sparse way, so one can say,
> natural selection naturally designs words that are good for efficient
> communication.

What's efficient is determined by the environment. Apes' hands are
efficient in forests but not in seas. Our linguistic environment has
been massively changing through the inventions of printing, internet,
etc., and we can question how well are natlangs keeping up with this
shift in media ecology at such an increasing rate.

Also crucial is the changing social structures & norms. How often do
we say "he or she" for the lack of a better pronoun to refer to the
increasingly common gender-irrelevant sets of individuals? Would you
expect a 'random natural' process to give birth to a good neuter to
replace this inefficient three-word phrase?


mu'o

--=20
You received this message because you are subscribed to the Google Groups "=
lojban" group.
To post to this group, send email to lojban@googlegroups.com.
To unsubscribe from this group, send email to lojban+unsubscribe@googlegrou=
ps.com.
For more options, visit this group at http://groups.google.com/group/lojban=
?hl=3Den.