Technical Artifacts, Engineering Practice, and Emergence 279
many risks that cannot be identifi ed by following the functional decomposition that is
implicit within techniques such as FMECA [Failure Modes, Effects, and Criticality
Analysis].”
- The unexpectedness and/or unpredictability of emergent features. If emergent
features are unexpected or unpredictable, then it seems rather odd to try to design emergent
properties into systems, no matter how desirable they may be. Is it at all possible to design
such properties into a system and is it possible to prevent them from occurring? Of course
engineers are accustomed to unexpected and unpredictable behavior associated with the
things they design and produce. From a traditional engineering perspective, however, such
behavior is to be prevented as much as possible. Engineers are convinced they can avoid
unanticipated behavior if they have enough knowledge about the systems they design and
make, that is, given enough knowledge about the behavior of the systems’ parts and how
they are related. However, if emergent properties are by defi nition unexpected or unpre-
dictable, then this is impossible.
The reason these issues are of particular importance in traditional engineering practice is
that they pose a serious threat to what is called the “control paradigm”: under the condi-
tions of operation and use laid down in the design specifi cations, the behavior of a technical
system can be fully controlled by controlling the behavior of its constituent parts. Emergent
features endanger the control of engineering systems. Apparently, emergent system fea-
tures with causal powers of their own cannot be controlled through the causal powers of
the systems’ constituent parts. Functional decomposition is the tool engineers use to con-
struct the behavior of the overall system starting from the behavior of its parts, and it
allows them to control the behavior of the whole system by controlling the behavior of
these parts. It would be pointless in the context of designing emergent properties. In regard
to unanticipated behavior, if there is one thing that does not fi t well with the mind-set of
engineers, it is that the things they create display unexpected and unpredictable features
(certainly when the thing is used or operated within the design specifi cations). It does not
matter so much whether the emergent features are desirable or undesirable; it is the lack
of control implied by unexpectedness and unpredictability that unsettles engineers. So it
seems that emergence and control are uncomfortable bedfellows. According to Buchli and
Santini (2005: 3), “there is a tradeoff between self-organization [and emergence] on one
hand and specifi cation or controllability on the other: If you increase the control over your
system you will suppress self-organization capabilities.”
This chapter does not address the issue of tensions between emergence and control in
complex engineering systems in a direct way. That would not only involve a comparison
of the various engineering notions of “emergence” that are in use but also require a detailed
account of the various kinds of complex engineering systems that are said to show emer-
gent properties (ranging from complex physical and technical systems to complex software
systems—particularly multi-agent-software systems—to sociotechnical systems). From
the point of view of engineering applications, many discussions about the benefi cial