*****  To join INSNA, visit  *****

> Loet and Ryan are commenting with admirable succinctness. :)
> How about this: Any relationship between two "units," from a 
> tension to a friendship, can be considered a "vector."  
> Whatever the content (relational modality)of the vector, its 
> degree of control over the actions of the units is a 
> variable.  The value of the variable runs from highly 
> cohesive (sticky--allowing no agency by the units) to highly "friable"
> (crumbly--allowing much agency by the units).  If we think of 
> it this way, we can compare networks on their degree of 
> cohesiveness.  

Dear Jeffrey, 

I would be inclined to think of the vectors as providing the variation and
the matrix as containing the structure ("eigenstructure"). The structure is
determining since selective. Variation can be random, but selection is
deterministic. In addition to eigenstructure at each moment of time, one
would have to define an eigen-dynamics over time. 

The eigenvectors and eigenfrequencies can be recombined by a self-organizing
system because it contains one more degree of freedom for the reflection. In
sum: a vector is a one-dimensional probability distribution, a matrix a
two-dimensional one, a system which develops over time would need to be
modeled as a three-dimensional probability distribution, and a
self-organizing system as a four-dimensional one. The entropy statistics of
probability distributions of more than one are not different from the simple
ones. Thus, one can move this forward. 

The cybernetic expectation is that the systems are constructed bottom-up,
but that control tends to emerge at the next-order level.

With best wishes, 


SOCNET is a service of INSNA, the professional association for social
network researchers ( To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.