278 Peter Kroes
properties is that their occurrence is often unexpected and unpredictable. Engineers would
not be engineers, however, if they did not try to harness these emergent properties for
human benefi t. Within an engineering context, the challenges that emergent phenomena
pose are primarily not intellectual—they are practical in nature. Understanding emergent
properties is one thing; manipulating them to solve practical problems is another. There
certainly appears to be a need to manipulate emergent features within engineering practice.
On the one hand, emergent phenomena in complex systems can have disastrous effects;
blackouts in electric energy supply systems are often claimed to be emergent features of
these systems.^3 On the other hand, they may be benefi cial with respect to certain desirable
properties of those systems; for instance, complex adaptive systems may be more robust
to changing conditions in the environment.^4 Depending upon whether the emergent fea-
tures are undesirable or desirable, engineers face the task of either trying to control or
avoid the emergent features or design them into systems. This is a branch of engineering
sometimes referred to as “emergence engineering” or “complexity engineering” (Potgieter
2004; Buchli and Santini 2005). However, is it at all possible to take emergence into
account and design systems in a predictable and controllable way, given that these emer-
gent properties are characterized as unexpected and unpredictable? Does the combination
of engineering and emergence make sense at all?
Discussions of emergence within engineering practice suffer from the same problem as
the ones mentioned here: there is much confusion about the meaning of the term emergence
and consequently about its implications for engineering practice. I make no attempt to
review the various interpretations given to emergence.^5 Instead let me point out three issues
from the various discussions about emergence that strike me as being of particular impor-
tance for engineering practice.
- The causal powers of emergent features. Emergent features without causal powers
of their own (often called “epiphenomena”) appear prima facie not to be very interesting
for engineering practice; they do not present new opportunities for causally infl uencing
the physical environment. Why bother trying to predict and control phenomena that have
no causal impact on the world? Emergent features with causal powers are more interesting
in this respect but confront engineers with the problem of how to control these causal
powers. - Emergence and functional decomposition. To tackle complicated design problems,
engineers deploy a divide-and-conquer strategy known as functional decomposition:
the function of the system as a whole is divided into subfunctions (and so on) that
are performed by the constituent parts of the system. This strategy does not seem to
work for systems functions that are based on emergent properties, since these properties
cannot be understood in terms of the properties of the parts of the system. As Johnson
(n.d.: 5) remarks, “The idea that there are properties of systems that cannot intrinsically
be understood in terms of lower level concepts seems entirely at odds with many
contemporary approaches to engineering. For example, this would suggest that there are