284 Peter Kroes
functions based on or realized by those emergent causal powers for use in technical
systems. If inductive reasoning leads to the conclusion that certain types of complex
systems show some kind of emergent causal power, it will be possible of course to intro-
duce functions based on this behavior into the design of a new system by designing that
new system such that it is a member of that type. However, this is more like imitation than
true design.^18 What is missing is an insight into how the overall function of the system is
realized through the subfunctions of its parts; in other words, what is missing is a func-
tional decomposition of the system that (deductively) explains how the overall function
results from a combination of all the subfunctions. But if the overall function is realized
by an emergent property, it is by defi nition impossible to come up with a functional
decomposition, since the emergent behavior (corresponding to the overall function) cannot
be reduced to the behavior of its parts (corresponding to their subfunctions) (Johnson n.d.;
Pavard and Dugdale 2000; Buchli and Santini 2005).
From an engineering point of view, the control of emergent causal powers also raises
serious problems. If emergent causal powers cannot be ontologically reduced to the causal
powers of the parts of a technical system, then it is questionable whether these causal
powers can be controlled in the same way that the nonemergent causal powers of the
technical system as a whole are controlled. The latter are controlled by means of “local”
control parameters, that is, control parameters that affect the causal powers of the parts of
the system and the way these causal powers are combined. All the local control variables
together exhaust the possibilities for intervening, and thus controlling, a technical system.
So it seems that insofar as emergent causal properties can be controlled at all, they must
be controlled by local control variables. One form of control seems less problematic,
namely to switch on or off emergent causal powers by effectively changing the kind of
structure of the system (which is conceptually different from controlling the original
system) such that the resulting system no longer exhibits the emergent causal powers.
Recalling the three issues described in the introduction, we observe that the fi rst issue
involves emergence in an ontological sense. This form of emergence also touches upon
the second issue, as it is connected to ideas underlying the technique of functional decom-
position. Emergent causal powers prove to be incompatible with the control paradigm of
traditional engineering. The occurrence of emergent causal powers in technical systems,
however, is, if conceptually coherent at all, highly speculative.
16.3 Technical Functions and Epistemic Emergence
In an epistemic reading, notions such as “novel,” “qualitatively different,” and “reduced”
are to be interpreted in terms of relations between knowledge of emergent features and
knowledge of features of the emergence base. This implies that relations of epistemic
emergence “turn crucially on our abilities or inabilities to comprehend or explicate the