Hi all,
Communication topology-social network, is, I think a good example of what a
building can be in an architect view of macro structure. Autonomous agents
interacting in a given environment. The problem is some buildings
(=topologies) which imply certain social dynamics whether others not.
Question: a star topology (=a star blueprint of a building) really tends to
stagnate dynamics of people leaving in it? Please respond, this is not a
rhetorical question...
Best,
Ioannis
PS. A star topology is a network where all agents communicate via a central
one and only the central one communicates with all others directly...
-----Original Message-----
From: News and discussion about computer simulation in the social sciences
[mailto:[log in to unmask]] On Behalf Of Nuno David
Sent: Thursday, July 09, 2009 10:08 PM
To: [log in to unmask]
Subject: Re: [SIMSOC] [Fwd: Re: [SIMSOC] Models incorporating both,
micro->macro + macro->micro ?]
Hi Keith,
I agree that no simulation yet combines both micro->macro and macro->micro
simultaneously.
However, I think most of modellers doing social simulation agree this
happens in the real world. For me, this is not the reason behind the
problem.
Rather, the problem is computational.
Rather then combining micro->macro and macro->micro simultaneously what the
simulations don't do yet is combining micro->macro and emergent-macro->micro
simultaneously.
If we want that "simulation automatically develops explicit data structures
that capture the emergent macro level structures" then I assume there must
be some mechanism doing that, which must be programmed by someone. Micro
agents shoud be programmed to capture emergent macro level effects before
the agents ant the programmer themselves knew what to capture at the macro
level. This happens in the real world indeed, but how do we program that?
How to programme an agent to observe and conceptualize?
The interactions in the simulation create interesting dynamics and we are
indeed able to model and describe this dynamics by capturing micro and macro
descriptions, including relationships between micro and macro descriptions,
etc, which is very useful.
But this is not the same as automatically developing explicit data
structures that would capture emergent macro level structures which then
would have explicit causal effects at the micro level. However, the latter
is not an ideological or theoretical question at the social theory level, it
is rather at the computational level. Once we find that we can implement
micro->macro and emergent-macro->micro simultaneously, I think everbody
will do that (including the methodological individualists). But I dont know
how to do it.
best regards
Nuno David
----- Original Message -----
From: "Keith Sawyer" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Thursday, July 09, 2009 7:08 PM
Subject: [SIMSOC] [Fwd: Re: [SIMSOC] Models incorporating both, micro->macro
+ macro->micro ?]
>I enjoyed these comments about upwards and downwards causation in social
>systems. Sorry to take so long to reply, I have been traveling.
>
> I'm not aware of a more recent literature review, since my 2003 article,
> of sociological implications of agent based modeling. (By the way, my
> 2005 book SOCIAL EMERGENCE is a bit more recent and incorporates an
> updated version of the 2003 article.) I'm interested if anyone can
> suggest one?
>
> Yes, it's true I claim that no simulation combines both micro->macro and
> macro->micro simultaneously but here's what I'm referring to:
>
> 1. At the micro level, agents interact locally, and this eventually gives
> rise to the emergence of some macro level structures and/or entities.
>
> 2. The simulation then automatically develops explicit data structures
> that capture the emergent macro level structures. (These are not internal
> to agents, but are autonomous from them.)
>
> 3. These new data structures then have causal effects on the local agents
> and their interactions.
>
> It is step (2) that results in the most controversy/discussion, with many
> modelers saying, of course we don't do that because that doesn't happen in
> the real world. But in fact, many sociologists do believe that something
> like that is the case in the real world. They are sometimes referred to
> as sociological realists. It's primarily those who are methodological
> individualists by inclination who think that (2) is unnecessary (as I have
> argued in several other publications). Most if not all agent modelers are
> methodological individualists.
>
> The mechanism whereby step (3) happens has been the topic of a few of
> these posts; immergence/second-order emergence is one of the leading
> candidates for a mechanism. But that's not the only one; macro social
> structures and properties often have demonstrable causal effects on
> individuals even when those individuals are not consciously or explicitly
> aware of them.
>
> -------- Original Message --------
> Subject: Re: [SIMSOC] Models incorporating both, micro->macro +
> macro->micro ?
> Date: Wed, 17 Jun 2009 16:13:11 +0200
> From: Rosaria Conte <[log in to unmask]>
> Reply-To: Rosaria Conte <[log in to unmask]>
> To: [log in to unmask]
>
> I would say yes. See for example the results of the EMIL project
> (EMergence
> In the Loop, a FET-funded European project under the 6th FW,
> http://emil.istc.cnr.it/ ) where the simulated process of micro-to-macro
> emergence of norms is shown to include macro-to-micro processes in which
> norms get gradually "immerged" in agents' minds
>
> Rosaria Conte
>
> National Research Council,
> Institute of Cognitive Science and Technology,
> V. S. Martino della Battaglia, 44, 00185 Roma.
> LABSS (Laboratory of Agent Based Social Simulation):
> http://labss.istc.cnr.it
> http://www.istc.cnr.it/createhtml.php?nbr=70
>
> Il 17-06-2009 10:29, "Georg Holtz" <[log in to unmask]> ha
> scritto:
>
>> Sawyer (2003) relates multi-agent systems to the micro-macro link in
>> sociological theory and concludes that there are simulations that show
>> how macro-social phenomena emerge from individual action and such that
>> demonstrate that a change of macro structure (e.g. network topology,
>> size of a society, communication mechanism) changes the bottom-up
>> processes of micro-to-macro emergence. He then argues that no simulation
>> has combined both micro-to-macro and macro-to-micro processes
>> simultaneously (to avoid any misunderstandings, of course emergent macro
>> situations influence micro-to-macro through specifying the context in
>> which individuals act on the micro-level. In those cases agents are not
>> influenced by the overall macro situation but only by local
>> interactions. But this is not what Sawyer refers to, he is interested in
>> a direct causal role of the macro level).
>>
>> However, in my view Gilbert's extension of the Schelling model in which
>> patches are labelled "good" or "bad" places for the respective groups
>> according to the history of the patches would be a model which involves
>> both micro-to-macro and macro-to-micro causation (Gilbert 2002)?!
>>
>> What has happened with regards to that in the last 6 years? Is there
>> any good up-to-date review article available on that topic?
>>
>> Thanks & regards
>> Georg
>> ---
>> Sawyer (2003): Sawyer, Keith, "Artificial societies - Multiagent Systems
>> and the Micro-Macro Link in Sociological Theory", Sociological Methods &
>> Research, Vol. 31, No.3, 2003
>> Gilbert (2002): Gilbert, N., "Varieties of Emergence", Transcript of the
>> introductory talk given at the Workshop on Agent 2002 Social Agents at
>> the Ecology, Exchange, and Evolution Conference 2002
>> (http://www.soc.surrey.ac.uk/staff/ngilbert/ngpub/paper148_NG.pdf)
>
> --
> R. Keith Sawyer
> Associate Professor
> Washington University
> Department of Education
> Campus Box 1183
> St. Louis, MO 63130
>
> www.keithsawyer.com
|