From: Joakim Almgren Gāndara (claw@lords.com)
Date: Fri Jan 19 2001 - 16:39:22 MST
First, my disclaimer:
Although it is one of my longest e-mails, this is not an actual essay,
that's just a categorization of the message. Thought I'd use the topic
system for once, just to point out that this e-mail is a personal thought,
not news or scientific facts. Stop me if <SL4 or if you've already discussed
this.
I don't know if anyone has ever attempted to think in Rap Lin Rie (also
known as Dutton speedwords) or any other shorthand system. I am merely
proposing some systems and languages that might optimize human or artificial
thoughts (note: in this text, I use "thought" to refer to conscious thoughts
that are somehow verbalized by our "internal narrator"). I will focus on
human thought, since I would expect that an AI created by humans would be
rather anthropomorphic.
...
I have often cursed my brain for constantly thinking my thoughts "out loud".
I often feel that the verbalizing of thoughts is a bottleneck, that the
brain narrates after I've expressed the thoughts subconsciously, forcing me
to wait until it's done narrating before I can continue thinking. Sometimes
I get stuck in the middle of a thought simply because I can't find the word
I need or because the English word eludes me. Yes, Swedish is my first
language, but I often think and dream in English.
I've tried to bypass this problem by concentrating on not verbalizing my
thoughts or by just interrupting the narrator and jump on to the next
thought, "cut to the chase" so to speak. I find that I can do without the
internal narrator for short whiles, but that I tend to lose my train of
thought and not reach any useful conclusion; my thoughts only fan out and
scatter wildly, a mildly Lovecraftesque sensation that reminds me of my
fever-induced nightmares and states of mind that seem frightening for no
apparent reason. It seems to me that my internal narrator has three main
purposes: 1) It helps me record and categorize my thoughts for future
reference, 2) filters out irrelevant thoughts, conjectures and memories, and
3) tries to direct the flow of thoughts towards some kind of conclusion.
(Perhaps it only performs function 1 or 2, and the other two follow from one
of those? Any thoughts?)
My conclusion is that a narrator is needed in order to organize
thoughts, and that to optimize the speed of conscious thought we need to
either optimize our language(s) or invent a shorthand language designed
specifically for thinking. There are a number of shorthand languages already
available; I will use Rap Lin Rie as an example here.
For those of you who haven't heard of RLR, here's a sentence from the
Gettysburg Address to give you an idea of how it works: "T e alge oki & oki
k w debi fa c." ("It is altogether fitting and proper that we should do
this.") Note that there is little nuance in RLR ("oki & oki") in order to
optimize for size and speed. Dutton obviously didn't care much about
aesthetics when he designed the language, he tried to construct an
international shorthand language. However useful it might be in writing, it
is hardly useful in speech. (How would you pronounce &? "Et"? That means
"little" in RLR. What about "w"?) This makes it rather useless as shorthand
thought language as well, unless one can learn to think using only
typographical representations of words. Since I think it's possible, RLR
might be a good candidate.
Other existing languages that might be considered are Chinese or
Japanese or any non-inflected language. Sadly, the many nearly identical
words have resulted in awkward constructions to specify which meaning a
certain word has in a certain instance, such as "look-see" in some parts of
China (this might be obsolete). Anyway, most natural languages tend to have
the same average speed and size, which rules out most currently spoken
languages as thought optimizers.
I've been toying with the idea of a parallell language, a language that
expresses many things in parallell. For example, I could be talking to John
about the weather while I talk to Jane about the mad cow disease. Sort of li
ke how DNA can be read in six different ways. Imagine the intricate
double-meanings and puns that would be possible in such a language! However,
I have no idea how to construct a parallell language. The closest I've come
is a kind of yin-yang version of Morse code, where you can either interpret
the message as usual or invert it, interpreting the pauses as beeps and the
beeps as pauses. For example, Long Beep (Short Pause) Long Beep also means
(Long Pause) Short Beep (Long Pause). However, this is obviously not a
parallell language; it's just two multiplexed channels.
Personally, I think it's pure lunacy to use only one of our five (six?)
senses when representing thoughts internally. Why not think in images,
odours, sounds, etc, combined? Why not construct a language that consists of
all the senses, and multiply current thinking speed with a factor of five
(six?)? In such a language, each concept is represented by a sensation. It
would probably be impractical, and it wouldn't help me at all, since my
problem is remembering certain words; in this language, the vocabulary would
be huge, since the idea is to have one sensation for every single concept.
I'd love to read any thoughts on this. I think that this is relevant to SL4
because I feel that to achieve transhuman intelligence we have to understand
the limitations of the mind and try to overcome them. Has anyone else given
this matter some thought? Please reply -- privately, if the subject is
buzzed. (Always the optimist.)
- Joaquim Gāndara
. http://www.ite.mh.se/~joaal98/
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:35 MDT