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

I just want to follow up on Blyden's post below, because so much of my
own work deals with the differences and distinctions between an
"object-oriented" ontology and one that is "relational."

An object-oriented ontology views objects (things) as primary, and
relationships as outside of the "boundaries" of that object.
Relationships are something and object "enters into."

A relational ontology views relationships as fundamentally constitutive
of what an entity actually IS.  Relationships are not something "entered
into" because the entity has no real properties outside of its existence
"in relation" to everything else.

I was led to relational ontology through studying ecosystems, where
biological networks are intrinsic to what a species "IS" in nature, its
very being.  Niches in ecology (which are relationships like
predator/prey) exist PRIOR TO the species' that come along to fill
them.  The relationships persist, but the relata change.

It was no large effort to realize that this applies to social networks
as well.

In political science, we use "Constructivism" which posits a fundamental
co-constitution between structures and agents, and gets us a lot further
along in recognizing the relationships, but falls short of true
relationality, because it focuses on structure and agents as objects and
not on the relations in the system.

Eventually, I reached a win/win solution with all of this, which is to
visualize the whole debate as squares on a chessboard.  Are they white
squares on a black board, or black on white?  Does it matter?  When we
choose to, we can focus on the "nodes" in a network as long as we are
clear that boundaries and identity are interpenetrated and permeable.
At other times, we can focus on the "relations" in a network, as long as
we accept that those relations evolve over time and can be extrinsic or
intrinsice to the entities in those relationships.

Would really love some feedback on all this. :-)


[log in to unmask]
The Universe is made up of stories, not atoms.

On Wed, 2004-08-11 at 12:50, Blyden B. Potts wrote:
> "I would add to this by suggesting that a relation cannot be defined without
> also defining a node....  A relation is a means to specify a specific
> association between *particular* nodes."
> Or maybe nodes are ways of defining entities from a particular set of
> relationships?
> "A node is a distinct and bounded measurable entity (often referred to as
> ego or an agent or an actor). Nodes can be people, books, businesses,
> countries, you name it. If it's discrete and non-trivial then it works.
> Oceans, for example, wouldn't work by this defition. They are continuous,
> not contiguous."
> The "boundaries" that lead us to see people, books, businesses, countries,
> etc as "distinct" or "discrete" might be just as subjective as the notions
> that lead us to consider the oceans "continuous".
> How do you even know that a boundary exists except by virtue of a
> discontinuity in relationships? You find the edge of your body by
> recognizing that on one side of it a series of relationships hold (e.g.
> continuity of flesh for example) that do not continue across the boundary,
> no?
> Consider a current in the above mentioned oceans. Where does one current end
> and another begin? What are the nodes that this relationship connects?
> In the modern age it seems increasingly that the old, assumed boundaries
> that defined a business or book or polity (at least) as discrete are no
> longer valid to assume.
> Can there be a definition of any THING (i.e. a node) without reference to
> how that thing relates to other things?  Including most notably a symbolic
> representation of that thing? If so, it would seem that relationships
> logically precede things or at least the identity of things as things.
> It seems to me presumptive to privilege nodes over relationships.
> Blyden Potts

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.