282 Peter Kroes
time.^11 But this behavior is not suffi cient for realizing the function. Note that this behavior
is not an emergent property of the physical construction; it is generally considered to be
realized by the physical construction and therefore to be ontologically reducible to the
behavior of its parts without any problems. However, that physical behavior by itself does
not endow the physical construction with the function of measuring time—with the prop-
erty of being a clock. For that, another necessary condition has to be fulfi lled, namely that
the physical construction is embedded in intentional (social) practices (in which it was
designed as a clock and used as a clock, etc.).^12 The foregoing means that the functional
property of being a clock is ontologically a relational property, which involves as relata a
physical object, with the right physical capacities or behavior, and intentional actions or
intentional states of a certain kind.^13
So the physical structure by itself does not realize the technical function, which means
that the second condition for functions to be ontologically emergent properties is also
satisfi ed. Therefore, relative to its physical emergence base, a technical function is onto-
logically emergent. One might object to this conclusion on the grounds that it is based on
an inappropriate choice of the emergence base. Suppose that the emergence base is
enlarged to consist not only of a physical construction but also of actions that occur in
intentional practices in which this physical construction is designed as, used as, and so
on, a clock. A strong case could be made for the claim that relative to such an enlarged
emergence base, technical functions are not ontologically emergent. However, the strategy
of enlarging the emergence base runs the risk of trivializing the whole notion of “emer-
gence” (it may be possible to turn any property into an ontologically nonemergent one by
an appropriate choice of emergence base).
These considerations make clear that the question of whether functions of relatively
simple technical artifacts are ontologically emergent properties or not, in the fi rst inter-
pretation described earlier, is not a straightforward matter. Philosophers may worry about
this form of ontological emergence of functional properties; engineers usually do not care
very much about it. From the point of view of the control paradigm of engineering practice,
the second interpretation, (b), which relates emergence to new causal powers, appears to
be much more relevant. So let us now turn our attention to that possibility.
To begin with, note that the interpretation of the fi rst condition for emergence (novel,
qualitatively different features) as new causal powers that go beyond the causal powers of
the system’s parts already takes care of the second (nonreduction) condition if we take the
meaning of “going beyond” to be the same as “not reducible to” (which we do from now
on). In contrast to the fi rst interpretation of the novel, qualitatively different condition, this
interpretation is rather strong: “emergence” in this sense implies that technical artifacts as
a whole have new causal powers in comparison to the causal powers of the parts of which
they are constituted.
We have already observed that systemic properties of technical artifacts without causal
powers of their own are not of much interest for engineers. They do not offer new oppor-