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

When I wrote the abstract that began a previous thread with this subject, I
was responding to Stephen Borgatti and Virginie Lopez-Kidwell's "Network
Theory" article in the SAGE Handbook of Social Network Analysis and
beginning to consider the proposition that besides network flow and network
architecture theories, SNA might also require a third, as yet undeveloped,
body of network chemistry theory. Among the responses was one from Barry
Wellman, who directed my attention to Barry Wellman and S.D. Berkowitz,
ed., Social Structures: A Network Approach. Having secured a copy of that
book, I was delighted to find in it a revised version of Ronald Breiger's
classic article "The duality of persons and groups," which first appeared
in Social Forces in 1974. I had seen frequent references to this article in
later work; but now I held it in my hands and could read it.

What I find in Breiger appears to me to be a set of compelling ideas that
have shaped the analysis of two-mode networks ever since the article was
published. Here let me begin with a quote from the opening paragraph.

"Individuals come together (or metaphorically 'intersect' one another)
within groups, which are collectivities based on the shared interest,
personal affinities or ascribed status of members who participate regularly
in collective activities. At the same time, the particular patterning of an
individual's affiliations (or the "intersection" of groups within the
person) defines his or her points of reference and (at least partly)
determines his or her individuality."

As I continue to read the article I note that while the possibility that
the individual members of groups may have different interests, affinities
or ascribed statuses and that individuality may be more than the sum of
overlapping group identities are briefly acknowledged, the translation of
the metaphor into mathematical matrices eliminates these concerns. Both
individuals and groups are treated as atomic units. They may be joined by
different types of relationships, giving rise to multi-relational networks;
but the impact of those relations on the internal structure of groups and
on processes of group formation and and their implications for group
activity remain unexplored. This, however, is precisely the molecular level
at which network chemistry operates and a level of analysis essential, I
would argue, for the understanding of project teams.

I am aware that Breiger was writing in the early 1970s, and there may be
whole libraries of more recent work that address these issues of which I am
ignorant. If so, I would welcome pointers to the work in question.
Meanwhile, however, let me review once again why project teams are a
stumbling block for current forms of network analysis and simulation.

   1. Project teams are not composed of people who happen to like each
   other, or seek information or more direct assistance from each other.
   Project teams are constructed with particular purposes in mind.
   2. Project teams assemble people who are not like each other except in
   the very loose sense that they belong to a universe of individuals with
   skills that may be needed for the project. Homophily is not a factor in
   decision making.
   3. Whether individuals are selected for project teams depends on
   multiple factors.

   - Necessary skills -- the skills required to play their assigned role.
   - Awareness--those in charge of recruiting team members may not be aware
   that an individual is qualified.
   - Availability--Individuals who are known to possess the necessary skills
   may be barred other commitments or group boundaries from participating in
   the team.
   - Previous experience -- candidates may have worked together before.
   Successful collaboration may increase their chances of working together
   again. Previous failures or quarrels may reduce the likelihood of another

In the case of the award-winning Japanese creatives who make up the
networks that I am studying, we can add additional considerations. Which
skills are required depends on the kind of ad a creative team is producing.
TV commercials, for example, require different skill sets from print
advertising. Team size may be directly affected. Teams that make TV
commercials are, on average, twice as large as those that produce print
ads. To which I must add an historic shift in the nature of the industry
itself: A major factor affecting the composition of the networks I study is
changes in the relative proportions of different kinds of ads, as TV became
the dominant advertising medium and print media declined. Shifting
proportions of different-sized teams directly affect network structure.
Shifting skill requirements shift the proportions of the various types of
relationships (creative roles) that link team members to teams. How these
changes affect the internal dynamics of teams remains an open question.

As always I am looking for comparable studies or comments that will
stimulate fresh thinking. I look forward to hearing from anyone who is
willing to chime in.

John McCreery
The Word Works, Ltd., Yokohama, JAPAN
Tel. +81-45-314-9324
[log in to unmask]

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.