|
|
To Owen: Here it is in a gross of words. Since probably the beginning of science (and even before) we have tended to think of higher level things as being composed of lower level things in a way that allows us to link the lower level to the higher level. That's only natural. Everything that has "components" is understood that way. However (as you and every other computer scientist or software developer on this list knows) abstract data types (and most other abstractly defined higher level things) may be implemented in any of a number of ways. The implementation is not relevant as long as the specification is satisfied. That's one of the insights computer science has bought to the world. I argue that the way to understand emergence is through that same lens--a higher level abstraction that is implemented by lower level elements--but not composed of them the traditional reductionist sense.
-- Russ
On Wed, Apr 29, 2009 at 3:08 PM, Ted Carmichael <[hidden email]> wrote:
On Wed, Apr 29, 2009 at 1:45 AM, Russ Abbott <[hidden email]> wrote:
Sometimes it seems to me that people prefer to think of emergence as mysterious. It's not.
I agree. That's generally the problem I have when folks start talking about downward causation.
I had an interesting conversation with Mark Bedau at a conference last month. He said (and I'm paraphrasing) that any conceptual model is fine, as long as you find it useful in some way. Emergence is a surprising property of a system because it is behavior that is not explicitly accounted for in the micro rules, and so it is useful to talk about emergence.
It is also useful to talk about downward causation. I think the error is when folks take an emergent property from the "surprising" category and put it in the "mysterious" category, as if it can't be understood in terms of rules and interactions of the system's constituent parts.
It's useful to say: when a school of fish turn left, that behavior influences a single fish to also turn left. But we all know, from programming these models, that the single fish is really only influenced by his simple rules, and the behavior of the nearby individuals that he interacts with. (And he, of course, influences them in turn.)
When I program CAS models, I try to avoid programming any downward causation. I prefer to have emergent properties come from the simple rules, rather than explicitly control for it with macro-level rules. I think to do otherwise is tricky business, and it's very easy to artificially introduce behavior that you are trying to explain, thereby undermining your explanation.
Personally, I think even difficult concepts - such as consciousness, or intelligence - can be explained without resorting to "mysterious" downward causation. (Even though we're not there yet.) And even if it is sometimes more useful to use a conceptual model of consciousness that utilizes macro-level causation, I think it would be a mistake to assume that this is the "real" model, or the only model available to us.
Just my two cents.
Cheers,
Ted
On Tue, Apr 28, 2009 at 9:29 PM, Nicholas Thompson <[hidden email]> wrote:
I should have said that the properties of a, b, c and E are synchronic.
> From: russell standish <[hidden email]>
> Date: 4/29/2009 6:14:43 PM
> Subject: Re: [FRIAM] emergence, again
>
> On Tue, Apr 28, 2009 at 09:33:42PM -0600, Nicholas Thompson wrote:
> > Here
> > is the kind of problem that bothers philosophers: let a, b, and c
> > constitute macro-entity E and let the behavior of E. be controled by the
> > properties and intereactions of a, b and c. Now, let one of the
behaviors
> > of E to control the behavior of a, b, or c. Is there a problem here?
> >
> > Nick
> >
>
> No. It sounds like a perfectly reasonable way of building a control
> system. Should there be a problem?
>
> --
>
>
----------------------------------------------------------------------------
> Prof Russell Standish Phone 0425 253119 (mobile)
> Mathematics
> UNSW SYDNEY 2052 [hidden email]
> Australia http://www.hpcoders.com.au
>
----------------------------------------------------------------------------
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org
|
Locked
