Re: Non-linear / full-2d writing systems?

> I would imagine Sai is thinking in terms of something with at least that
> level of expressivity; but my understanding is that what he wants
> expressed is something more akin to "pure thought" and not related to any
> spoken (or once spoken or could be spoken) language, whether a natlang or
> conlang.
A brief reply on this point:

I see it as two separate questions, though quite related.

First is: what does 2d syntax (/ morphology / fusion?) buy you? I tend
to view this in terms of a) information density per stream (i.e., what
is this encoding, compared to its linear equivalent) and b) goodness
of fit (i.e., is 2d a more 'natural' [as in intuitive not as in
natlang] fit for expressing certain semantics, and therefore we
can/should offload those onto it).

Both of those come with the assumption that any extra data you're
encoding into the 2d-ness, somehow, will be removed from other sources
and replaced by something else. That is, if you view it as information
density by stream, I don't want the density to go down in any stream
just 'cause this one is going up, and I'd prefer it didn't just
reduplicate / parallel (as you would get by, say, using natlang words)
it, since that's wasteful. (Viz. my "principle of semantic
conservation", if you've read ODIL.)

Second:

I am not against relation to natlangs or spoken langs per se. I just
don't particularly care whether it is, and I suspect that copying them
would make for fundamental inefficiencies.

And yes, I have a general principle of wanting my conlang
implementations to mirror thought as closely as possible. An ideal
language would be telepathy, basically; the closer we can get to that
- e.g. by having the language be as accurate and "native" a
representation of internal thought processes as possible (conscious
and sub-), the better. I for one don't think linearly, within my head,
until it comes time to communicate it or encode it in some form (like
for memorization). I feel that this is a bottleneck, honestly, and I
kinda resent that I have to resort to such because I don't have better
tools at my disposal; it's like my a lot of my thought gets lost in
translation into "normal" language, because of inherent low-level
flaws (or more generously, "mismatches") in the system.

(I have some techniques for working around this - like creating
triggers that merely index an existing, non-symbolized thought or
thoughtweb, to be invoked at will / memory - but that's rather kludgy
and doesn't address the questions of external record and
communication.)

> True - a totem pole is really a 2d surface curved round a cylinder.
Not if depth has meaning.

> Absolutely - But (as I think Sai realized) I was being a bit "tongue in
> cheek" with the 3d suggestion. The 2d idea is complex enough for starters
> - so let's not pursue the 3d idea (we can leave that to conlangers of the
> 22nd century   :)
Aww.... but I want an ultraneato dynamic holographic-projection
writing system... *whine*

> > Whether they are read linearly or not, to me, is beside the point.  A
> > human reader, bound by time and limited mental capacity, would
> > necessarily process the information linearly in one way or another.
Distinction: they would necessarily process it *chunked*. Chunked !=
linear - viz. experiments like "word superiority effect" that
demonstrate parallel processing, at least at low levels. But yes,
there would necessarily be a linear sequence *of* chunks, for some
value of "linear". I suspect that it need not even be that, since it's
not a one-item-then-next thing - there's multithreading and whatnot
going on.

> It is the notion that something like a whole book could be thus written. I
> do find it very difficult to see how the _sense/ meaning_ of Charles
> Darwin's "The Origin of species" or Leo Tolstoy's "War and Peace" could be
> written so there is more than one way you could read it and there is no
> starting or ending element.
Darwin I could see, given that it's basically setting forth an idea
(or rather, a set of ideas) and then going through all the support /
analysis / whatnot - not a necessarily linear thing.

Novels, however, are tricky. The way we conceive of a "story" now is
rather directly tied to the concept of events-through-time; it's a
core element. Without it, you don't get the whole build-climax-release
'thang', and that seems to be crucial to our culture's concept of a
good story.

So for storytelling, you would either need some form of linearization
to enforce that, some other way to acheive the same effect (*), or an
entirely new paradigm of what makes for a good story.

* For example, having different values for "tension" - perhaps one
could have something that is good over the net as a whole, so that
there would be a progression of tension / knowledge / understanding as
one comprehends more and more of the network and gets a more
woven-together / digested understanding of the whole thing. I think
this would be possible, and quite probably has a direct parallel to
the progress of enlightenment in the meditation-practice sense:
growing understanding through multiple paths, and comprehension of
patterns and fundamentals. It's an interesting little problem.
"Non-trivial", though, in the scientific sense of the word. :-P

 - Sai

Replies