[From Bill Powers (931005.1745 MDT)]
Chuck Tucker (931005) --
Interesting letter to Babbie. I'd like to introduce a note of
caution about one statement:
... objects, tastes, action results, and all of what I have
mentioned above EMERGE OUT OF INTERACTION; something different
that cannot be predicted from knowing about the elements arises
from the interaction.
Instead of saying that interactions "cannot be predicted," I
would recommend saying only that predicting them may be
difficult. In fact there are many instances in which we know the
properties of individual components _and their pairwise effects
on each other_, and from just that knowledge can predict what
will emerge when the components are interconnected.
An obvious example is a control system in relation to some
specific variable in the environment. We specify the properties
of an input function, a comparator, an output function, and an
environmental feedback link, and state how each one is affected
by its predecessors and affects the components that follow it.
Given these specifications, we can predict the emergent behavior
of the whole system under arbitrary changes in disturbances and
in reference signals. The whole system, but not any one component
of it, is a control system and exhibits the behavior that only a
control system shows.
In principle, if the properties of all components of a system are
individually known, and if the effects of each component on its
neighbors are known, it is possible to write down a system of
equations, each one of which decribes only a small part of the
whole system, and solve them simultaneously to reveal the
behavior of the whole assembly. So emergent behavior can be
deduced from knowledge of the parts of a system. It's not true
that in principle it is impossible to predict what the whole will
do from knowledge of its parts.
In practice, of course, systems don't need to have very many
components, and the components don't need to depart very far from
simple ideal properties, to make an actual system analysis
unachievable by analytical mathematics. But that is a practical
matter, not one of principle. Emergence doesn't create any
mysterious new properties that are disconnected from those of the
components.
But there is one aspect to systems analysis that is a bit
mysterious, and that is the effect of _organization_. A given set
of components can be assembled into a large number of different
systems with different emergent properties: a radio, for example,
or a bomb timer. What makes the difference is not their
properties but the way they are connected. In a device like a
brain, where individual components occur in great numbers but
only a few different types, the emergent properties become
overwhelmingly determined by the connectivity of the components.
That sounds like a promising concept for sociology.
ยทยทยท
-------------------------------------------------------------
Best,
Bill P.