The Extended Composer - Computers and Creativity - page 185

Fig. 7.4 The Wundt curve.
As novelty increases,
gratification rises to a peak,
falling again as we move
towards more extreme
unfamiliarity
7.3.3 Intimacy
To enter into a meaningful and enduring relationship with a tool or creative part-
ner, we must secure a degree of trust in it: trust that its responses will have some
relevant correlation with our own, rather than it disregarding our inputs and behav-
ing completely autonomously; trust that we can gain an increasing understanding
of its behaviour over time, in order to learn and improve our interaction with it, ei-
ther through embodied (tacit, physical) or hermeneutic (explicit, neural) knowledge;
and, in the case of computational or human partners, trust that its activity will con-
tinue to generate interest through autonomous creative exploration. In other words,
the output of such a system should be novel, but not too novel; as represented by the
Wundt curve shown in Fig. 7.4 .
Creative interaction with generative systems is often premised on a duality,
wherein the computational system generates material and the human acts as a fitness
function, selecting or rejecting materials and arranging them into a final product.
This would be a tiresome process if the generated material varied widely from what
was required. Consistency of operation also improves the confidence of an artist in
the output of a generative system. Confidence and predictability in the system con-
tribute to the development of a partnership and, ultimately, to the productivity and
quality of the work.
Predictability aside, it is clear that all designed artifacts, including generative
systems, are biased by decisions made by their developers and by the materials and
processes they use. We must align our thinking with the patterns and prescribed
methods that underlie the design thinking of the system (Brown 2001 ). Understand-
ing these patterns is necessary to get the best out of the system.
For an effective partnership with a computational tool, we suggest that it is neces-
sary to accept such biases as characteristics, rather than errors to be fighting against.
Again, taking the analogy of a traditional musical instrument, good musicians learn
to work within the range of pitch, dynamics and polyphony of their instrument as
they develop their expressive capability with it.
A quite different difficulty lies in the material status of our tools. Magnusson
( 2009 ) argues that acoustic and digital instruments should be treated with categor-
ical difference, with implications for our ontological view of their interfaces. The
core of an acoustic instrument, he argues, lies in our embodied interaction with it,
realised through tacit non-conceptual knowledge built up through physical experi-
ence. A digital instrument, conversely, should be understood hermeneutically, with
its core lying in its inner symbolic architecture. Tangible user interfaces are “but
arbitrary peripherals of the instruments' core” (Magnusson 2009 ,p.1).