280 Peter Kroes
exploitation of emergent properties are still highly speculative. Instead we deal here with
the question whether, and in what sense, functions of simple, stand-alone technical arti-
facts, such as everyday household appliances, may be regarded as emergent, and we
explore the implications of various forms of emergence for the control paradigm. A deeper
insight into these issues may be of help in understanding emergence in more complex
technical systems.
Before we continue, let us heed Van Gulick’s advice (2001: 27) and pay attention to
our “key.” The following general characterization of emergence is our point of departure:
emergent features in (complex) systems are 1) novel, qualitatively different features in
comparison to the features of the system’s parts, which 2) cannot be reduced to the features
of those parts and their relations.^6 Of course we must explicate in more precise terms the
meaning of the notions “novel,” “qualitatively different,” and “reduced.” Following Van
Gulick (2001: 16 sq), we distinguish between a metaphysical/ontological and epistemic
reading of this characterization of emergence; the former concerns emergence with regard
to real-world items, the latter with regard to our representations of the world. Our discus-
sion of ontological emergence (section 16.2) starts with the question of whether the func-
tion of a technical artifact may be regarded as an ontologically emergent property with
respect to its physical structure. We argue that this is the case. We also consider the pos-
sibility of emergent phenomena in technical artifacts with causal powers of their own and
we analyze in more detail the challenge this would create for the control paradigm within
traditional engineering practice. Epistemic interpretations of emergence focus on cognitive
relations between our knowledge and representations of emergent features and properties
of the entities from the emergence base. Epistemically emergent properties have a direct
impact on engineering practice since they signify the limits of predictability and explana-
tion and therefore the limitations of the control of technical systems. Our discussion of
epistemic emergence (section 16.3) fi rst addresses the question of whether the function of
a technical artifact is an epistemically emergent property. We argue that for simple techni-
cal artifacts, the function is emergent relative to a physical knowledge base, but that a
knowledge base may be chosen such that the function is not emergent. Finally, we explore
the possible impact of weak and strong forms of epistemic emergence for the control
paradigm.
16.2 Technical Functions and Ontological Emergence
Our ontological reading of the general characterization of emergence takes into consider-
ation two interpretations of the fi rst condition for emergence (novel, qualitatively different
features), namely (a) properties that can be attributed sensibly to the system as a whole,
but not to the parts of which it is made up, and (b) new causal powers that go beyond the
causal powers of its parts.^7