Perceptions, including physical, may cause emotions,
or may not, or emotions may arise without physical
perception (e.g., you figure out a solution to some
problem, or remember something). The fact that
perceptions may cause emotions doesn't mean they are
equal.
Perceptions may also cause feelings - e.g., cold,
or itch - those are different from emotions.
Urges are things like hunger, lust, etc.
The difference between urges and emotions, IMO, is that
urges are born on a level of an organ/subsystem, and
express its needs to the rest of the individual; emotions
are born on the level of the individual, and _may_ be
expressed socially. Expression is not the emotion though.
Example: anger is an emotion; frowning is an expression.
The need for emotions, as I see it, arises in primitive
systems that are unable to figure out causes and effects
and plan their actions intelligently.
Suppose you want to program a rabbit for optimal behavior.
You could explain to it its intended role in evolution,
and ask it to choose appropriate behaviors in all life
situations. That may be pretty hard for a rabbit to do
consciously, as it doesn't have enough general intelligence;
hardwiring a complex analytical system into the rabbit's
brain is too difficult for the blind evolutionary mechanism.
The solution here would be to program "one-step smarts" into
the rabbit. If you see a wolf, run; if you see a fox, run;
if you see a hawk, hide; if you see a falling stone, jump.
The rules would be more compact and general if they are
grouped.
(wolf,fox,hawk,stone) -> [State "danger", bad]
[Everything OK] -> [State "happiness", good]
Now, you program into the rabbit a desire to avoid bad states
and achieve good ones, and appropriate reactions for each state,
and the job is done. The perception of each of these states by
the rabbit is an emotion. The perceptions of its causes, and
resulting behaviors are connected to emotions, but are already
a different matter.
It is really easy to program such one-state rules into a computer.
You can write a chess program that would "feel good" every time it
takes an opponent's piece, "scared" every time its piece may be taken,
"angry" when it happens, and "peaceful" in all other cases. You can
even evolve a population of such programs with a mix of predators,
cowards, and peaceful/aimless piece movers.
One thing you can't do with such "single-step intelligence", is to
develop a good chess program, as chess requires looking more then
one step ahead. And so do all other problems. So the best solution
would be to come up with a brain that can do better-than-one-step
situation analysis, equip it with understanding of the rules and
goals, and do away with this primitive emotion stuff.
This is how we program the computers. They are already beyond emotions.
With ourselves, it is different though; as a legacy system, the human
consciousness is a funny mix of an "inner rabbit" and new rational
thinking mechanisms. This defines our identity and allows us to have
a lot of fun (usually by misusing the mechanisms wired into us by
nature to optimize our behavior in the woods and monkey tribes), but
may not be the best way to program computers, or engineer the next
generation of intelligence.
If I ever get a chance to reprogram myself, I would put emotions
under complete control, so that, when I want, I could turn off the bad
ones, and feel the selected good ones. So that the emotions would
be used only for fun, and will not stand in the way of decision-making
processes, where they tend to mess things up, even when the goal of
decision-making is to optimize future emotions. I guess this is
something most people are trying to do with their emotions already.
--------------------------------------------------------------------------
Alexander Chislenko Home page: <http://www.lucifer.com/~sasha/home.html>
Firefly Website recommendations: <http://my.yahoo.com> ---> "Firefly"
--------------------------------------------------------------------------