At last I found JCB's "LOGLAN 1: A LOGICAL
LANGUAGE" (Revised Fourth Edition- 2008) and downloaded it
and have been reading the various introductions etc. with
interest.  After all, it was JCB who originally coined the
word Loglan, from which _loglang_ is obviously derived; so,
whatever may have happened to the word Loglan since, going
back to the source, so to speak, seemed to me a sensible
thing to do.

Over the weekend I have been musing on earlier replies from
And, Logan and Alex and on what I've been reading from JCB's
book in between watching Rugby World Cup quarter finals, and
at long last yesterday evening began to type a reply, which
I intended to continue today - but Logan's email makes now
makes that a bit unnecessary    :)

On 18/10/2015 19:53, Logan Kearsley wrote:
> On 18 October 2015 at 05:31, R A Brown wrote: [...]
>> What then do I say on my webpage?

> Given And's admission* that his definition "should be
> considered a prototype definition", not a strict
> is-or-is-not test, I have come up with a new synthesis
> that I think takes adequate account of empirical evidence
> in usage as well as the opinions expressed so far. And in
> his defence, it was something Gleki said that triggered
> this condensation of thought; but, I could not figure a
> way to respond to that directly without further derailing
> things in a discussion of Lojban in which I am not
> qualified to participate.

Quite so - this thread was never intended to be about what
Lojban is and is not nor, indeed, about whether any
individually named loglang achieves or does not achieve

> So, here's my newest proposal (it's a bit long, but I
> don't think that can be helped):

'Tis a bit long - longer than I was hoping for something to
put on my webpage.  Time, perhaps, to bring out and dust off
the précis skills we had to learn at school for the old O
Level English exam    :)

> A 'loglang' is a language which is based by its creator
> on some system of formal logic -

Amen! I think a reference to formal logic is essential. Is
not that why we have the Log- of Loglan and Loj- of Lojban?
  Loglan was conceived on that basis around Christmas of 1955:
""At the beginning of Christmas Holidays, 1955, I sat down
before a bright fire to commence what I hoped would be a
short paper on the possibility of testing the social
psychological implications of the Sapir-Whorf hypothesis. I
meant to proceed by showing that the construction of a tiny
model language, with a grammar borrowed from the rules of
modern logic ..... "

> in practice, typically first order predicate logic-

Yes, indeed.

> A direct result of this is that all such utterances are
> intended to be completely *unambiguous*, at least at the
> level of syntax and above, though lexical ambiguity
> (polysemy) and vagueness may still exist.

Yes, I found that JCB devotes a lot of space to emphasizing
that the fourth edition of Loglan 1 is 100% unambiguous at
the syntactic level.  He does admit at the lexical level
there may room for metaphor and ambiguity. Personally I do
not see how that can be avoided if the language is going to
be used by humans.  we have a fondness for metaphor and for
shifting the meanings of words.

But at the syntactic level he was insistent there must be no

> Thus, the term 'loglang' refers not to a strict binary
> classification,

Is any -lang classification a strict binary classification?

> but to a scale- a language which enforces perfect
> unambiguity of logical form at larger scale or over
> broader domains is "more of a loglang" than one that
> does not. A 'pure' loglang is one in which *all*
> utterances at

I'm not sure I would want to get into the question of 'pure'
and 'impure' loglangs on my webpage. I'll read this snipped
bit again (and probably again) carefully.  It may be here
that the précis knife comes into play   ;)

> Commentary: As an example to demonstrate how a language

I think I need also to consider the example more closely.

> ... and a loglang because it enforces complete logical
> unambiguity in each clause- but not a pure loglang
> because it admits ambiguity above the level of an
> individual clause, and not total or prototypical because
> it cannot encode *arbitrarily complex* logical structure
> without ambiguity.

Yes, the arbitrarily complex logical structure probably
should figure in a definition.

"Ein Kopf, der auf seine eigenen Kosten denkt,
wird immer Eingriffe in die Sprache thun."
[J.G. Hamann, 1760]
"A mind that thinks at its own expense
will always interfere with language".