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

On Fri, May 06, 2005 at 07:25:03PM +0100, Ray Brown wrote:
[...]
> Non-liner/full-2d writing systems?
>
> _All_ writing systems surely have to be fully 2d otherwise we'd not be
> able to see the darn things? Sure, things like radio and electronic
> transmissions are not 2d - but *writing*?
I think what is meant is a writing system that is not confined to a
single dimension in its extension. Sure, the letters we write are 2D
curves and lines, but they represent atomic units which are composed
linearly to build sentences and paragraphs. They may be laid out on a
2D page, but they are read linearly: there is a beginning and an end
and only one path between the two.


> Now - a 3d writing system would be interesting   ;)
If you meant writing in 3D characters, that'd be right up the alley of
a conlang spoken by 4D beings. :-)


[...]
> On Wednesday, May 4, 2005, at 07:47 , Sai Emrys wrote:
> [snip]
> >By "full-2d", what I mean is something akin to a thoughtweb - rather
> >than a serialized / linear sequence of characters, it would
> >interconnect concepts in 2d (or greater) space. For example, copular
> >forms - equation, adjectives, subset, etc. - could probably all be
> >written by connecting two (symbolic) ideas with a line of a particular
> >kind, representing their relation.
>
> That sounds very much like the 'box analysis' we used to do in English
> lessons way back in the 1950s. 'Twas certainly 2d - with 'boxes' on
> different levels (the positioning was important) and the use of different
> types of connecting lines. I guess if one replaced the boxes enclosing
> English words with Chinese logographs it would be something like you have
> in mind above.
I think the idea goes beyond that: take a look at Pinuyo if you
haven't already --- it expresses complex concepts by
arbitrarily-complex layouts of pictograms, in 2D. You can potentially
write an extremely complex "sentence" in Pinuyo, involving nested
boxes of every sort at every point, and it will still constitute a
"single sentence". This can obviously be linearized (as with anything
else), but it is much easier to "read" in its 2D form.

Also, within a single Pinuyo "sentence", there is no fixed point at
which one must begin to read, and no fixed endpoint at which one must
end. There can very well be multiple paths from any chosen point to
any other point. So this writing has a non-linear (or should I say
"super-linear"?) component to it that is lost when you translate it
into a linear language.


[...]
> But whichever approach is used, one cannot, as far I can see, avoid some
> linearity/serialization.
> Even if whole sentences are depicted as a single symbol construct
> (glyph/frame/ or whatever) the sentences will surely follow sequentially
> or linearly. Even if complete paragraphs are written as single symbol
> construct, the paragraphs themselves will be linear - or if comlete
> chapters or - the mind boggles - a complete book is written as a single
> symbol construct* cannot surely dispense with linearity?
But see, this is the very barrier we are trying to overcome here: the
concept of a "sentence" or a sequence of paragraphs is inherently
linear. This is something very ingrained in us because existing human
languages are essentially linear.

But the point is, it doesn't *have* to be linear in order to be
understandable. Take for example, the description I gave in my post a
few days ago, of a "writing" on the wall which describes the story of
a hero who slays a beast:

The entire story is written as a very large, and very complex 2D
diagram comprising of interconnected parts. Each part describes a
different aspect of the story: one part may describe the hero's
ancestry, another part his past achievements, another part the beast
and its origins, yet another part the legendary weapon the hero uses
to slay the beast. Each of these parts are further decomposed into
sub-parts, and those sub-parts into their respective sub-sub-parts.
The "main event" of the story is represented as an interconnection
between the part of the diagram describing the hero, the part of the
diagram describing the weapon, and the part of the diagram describing
the beast.

At the finest level of detail, the diagram consists of symbols
representing verbs, nouns, etc., interconnected by various types of
lines indicating various grammatical relationships: agent, instrument,
patient, adjective, and so forth. These interconnected pieces are then
treated as larger units that are further interconnected, and so forth,
until you reach the highest level where the diagram has three parts
linked together by the main event of the story, the slaying of the
beast by the hero.

The underlying idea is that this diagram is a non-linear, connected
graph (in the mathematical sense of nodes and edges). It can be read
in one of a LARGE number of ways (equal to the number of unique vertex
cover paths in the graph). In fact, it doesn't even have to be read
entirely; one could, for example, simply pick out the main event,
consisting of the symbol representing the hero, the symbol
representing the beast, and the symbol representing the weapon,
each connected to the symbol representing the slaying.

Or one could pick out a sub-part of the diagram, which need not
include any character of the main event. For example, the hero may
have a notorious ancestor, represented, say, by some symbol X. X is
connected to H (the symbol representing the hero) by a particular line
that represents ancestry. X may be connected to another symbol, say Y,
which is the symbol representing his wife. Y can then be connected to
C, which is a gift given to her by some person represented by the
symbol D. One could, for example, read only the part of the diagram
that represents the giving of the gift to C by D, which would be
represented as a connection between C and D, mediated by a symbol
representing the act of giving.

This gift, G, could very well be related to another, distant part of
the diagram, say, to K, the former keeper of the beast: e.g., G is
connected to K by a symbol representing stealing --- G was stolen by
K. So this forms another sub-part of the diagram that does not
directly relate to the story, but is an interesting detail.

So in reading the story, one could start anywhere, say G, and traverse
the diagram in any order, and end anywhere one pleases. One does not
even have to traverse the main event of the story. But if one reads
the entire diagram, one gets a very detailed, very elaborate
description of the story. It transcends linear writing in that the
writer is not obliged to present the events in *any* particular order:
he presents the entire story with all of its detail, perhaps bringing
the important events to the forefront by employing, say, larger
symbols for the important events, and smaller writing for the obscure
details that fan-readers would love to know about. It is then up to
the reader how much detail he wants to read. One could even read the
diagram many times, at varying levels of detail, depending on one's
interest.


> *Whether the symbol construct maps to the spoken version of the paragraph/
> book or whether it rather represents its ideas & concepts is beside the
> point as far as my question is concerned.
Yes, what decides whether the representation is linear or non-linear
is in the explicit interconnections between symbols.

One could write the same story I describe above by devoting many long
chapters to elaborating on all the details of the various characters
in story. The reader could read through everything, and get the same
information about the interrelationships between the various entities
described in the story. However, the text of this form of the story is
linear, because its symbols must be read from start to end in exactly
one sequential order.

But in my version of it, the symbols are explicitly interconnected in
a non-linear fashion.


[...]
> As I said in the previous discussion arguing about its possibility or
> otherwise is not going get us very far IMO. The best thing surely is to
> attempt to do it!
[...]

I think I might try to attempt building a small-scale diagram over the
weekend to illustrate my hero-slays-beast example. I'll probably use a
rough, circled-English-words + lines notation for it, just to explore
the idea. A "true" writing designed for this type of notation will, of
course, employ symbols and styles of representing interconnections in
more idiomatic ways.


T

--
Verbing weirds language. -- Calvin (& Hobbes)