IEEE Technology and Society Magazine - September 2017 - 58

The form of governance that any of these organizations chooses is up to the organization itself. Organization governance choices range from strict dictatorship,
to representational democracy, to full involvement of all
participants in the decision-making processes. The form
of governance is, among other things, defined in the
constitutional choices made within the organization [12].
These choices define rules that, e.g., determine who is
allowed to make which collective choices affecting the
operational level. Constitutional choices thus determine,
e.g., how membership in an organization is decided and
which individuals are responsible for such decisions.
Such rules can be formalized as norms. If these norms
are the result of a collective choice, they should also be
"enforceable against non-conforming agents" [12].
If such procedures are applied in socio-technical systems, a host of potential problems arise. One of the more
pressing concerns is how to get everyone to adhere to
these new rules. If agents can make a normative choice
[4], then they can refuse any changes in the norms. That
this can be an issue has been recognized in a field of
study at the intersection of behavioral and social sciences, economy, and psychology that is devoted to understanding how changes in human organizations in the
workplace work. This field, change management [9], has
yielded numerous insights into the motivation of individuals to participate in a change, in their resistance, nonmonetary incentives, and how to plan, enact, and monitor
complex and far-reaching changes. In particular, this field
regards situations in which enforcement is either not possible or not desirable and, in contrast, where resistance is
embraced in the change process [30]. These insights are
interesting since the multi-agent systems and self-organizing systems community to date prefer a view of change
that is purely mechanistic, assuming purely rational
agents in the game theory sense of the word, and no
resistance to change if incentives are correct.
This paper is an attempt to outline the connections
between participatory self-organizing socio-technical systems and change management and to show alternatives
to this purely mechanistic view of change. In particular,
this paper proposes that the field of change management can lend insights into the technical solutions we
choose to ensure that changes within a socio-technical
system can be enacted by the social part, supported by
the technical part, and propagated in the system in a
way that embraces resistance and minimizes the need
for enforcement. We argue that this way, participation
becomes a real possibility rather than an afterthought.

Change, Incentives, and Resistance
The predominant approach to addressing issues of
change and how agents react to them in multi-agent systems research is to apply game theoretic concepts.

Game theory suggests that, should they be presented
with the "right" incentive structure, agents will happily
change and adopt any kind of bahavior that promises
the best pay-off. The trick for a system designer is to set
the incentives in a way that the individual bahavior optimizes for the intended overall outcome. This notion is
leveraged in technical systems in the approach of "mechanism design" [18], [20]. There are assumptions inherent
in this approach, however:
1) There is an assumption that all relationships and all
interactions within a system can be quantified and
rewarded by appropriate incentives that, in turn,
benefit the agents in some way in the form of "utility." Many examples of applied game theory literature, in particular in the domain of multi-agent systems
and mechanism design, suggest that incentives take
the form of monetary rewards. The advantage of
this approach is that the utilities of different agents
become comparable and it is very easy to find a
ranking of action alternatives that maximizes the
utilities within the entire agent population. In social
systems, however, incentive can take less tangible
forms such as "social capital" [21]. Many systems
also appeal to the collecting instinct of the human
users by handing out "rewards" for achievements
that bear no physical value but are rather valuable
with regard to the status of the owner within the
limited scope of that particular system (as discussed,
e.g., in [7], or in [24] in this issue). This means that
utilities become extremely subjective. The resulting
heterogeneity also makes attempts to analyze equilibria much more difficult.
2) Individual agent payoff and overall societal payoff
are supposed to be in a direct relationship where the
maximization of one maximizes the other. This leads
to situations such as the tragedy of the commons
[8] if this relationship does not exist. A healthy and
wealthy society that is characterized by a lack of
negative externalities and the long-term individual
payoff this yields is usually not regarded in individual
payoff calculation.
3) The agents are assumed to have full knowledge of
the expected incentives, of the different action alternatives they have, and of the long-term impact of
their choices.
4) Each of these assumptions is critical and each of
them is difficult to uphold in a complex socio-technical system. What happens if even one of them is
challenged? A possible consequence would be that
the agents in the system would resist a change even
though it would be rational to adopt it, and would
either try to maintain the bahavior they have applied
successfully so far or adopt a different bahavior that
they believe to be the one with the highest personal

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

SEPTEMBER 2017

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - September 2017