Information Technology Reference
In-Depth Information
In the following sections, I will discuss how the notions of generative and adaptive
creativity apply in various instances associated with human social behaviour. In
doing so I will draw on concepts from the social sciences which I believe can enrich
the conceptual foundations of computational creativity. The discussion is geared
towards arts-based computational creativity, but will draw from scenarios outside of
the arts.
14.3 Generative and Adaptive Creativity in the Arts, in Humans,
Human Groups and in Silico
The generative/adaptive framework can be applied to the way we characterise the
creativity of artificial systems in the poorly understood domain of human artistic
behaviour. Despite some clear achievements in arts-based computational creativity,
success has been marred by the challenges of evaluating the products of creative
systems in a meaningful way (Pearce and Wiggins
2001
, Greenfield
2006
, McCor-
mack
2008
). We struggle to disambiguate the creative input of the system's designer
from the creative output of the system itself (Greenfield
2006
), and we also struggle
to establish valid contexts for either subjective or objective judgement of creative
outputs. This ambiguity has made arts-based computational creativity particularly
resistant to the bootstrapping of future developments from previous success. It is
hard to make informed decisions about the relative virtues of different computa-
tional methods (genetic algorithms, neural networks, etc.) because arts-based com-
putational creativity research lives mainly in the lab, with many additional steps
required to get from this to the real field of human artistic activity. This suggests the
need for a richer characterisation of creativity, including the role of social dynamics
as a creative system, as well as the role of creative individuals within that system.
Above all, in labs, goals are implicitly defined for arts-based computationally cre-
ative systems which have ill-defined parallels in the world of human artistic goals.
A person has a flexible and creative relationship with the value of the cultural arte-
facts he creates and is surrounded by, for example by being free
not
to be artistically
creative. As currently conceived, an arts-based computationally creative system is,
by contrast, a tool with a function assigned to it. It can only be adaptively creative
within the limits imposed by its given function (it must make art). Functionality is a
desirable property of all manmade systems. An artificial creative system is expected
to do more than generative creativity (an end in itself), because it is required to pro-
duce outputs which are not just novel but have some externally established value.
As such, arts-based computationally creative systems target a novel niche, in terms
of their social embeddedness, which is yet to be clearly characterised. There is a
need to establish a discourse that properly recognises this niche.
What, then, is the nature of natural creative systems in terms of function? To
determine what kind of status a creative tool could take, it is important to look at
human creativity not only at the individual level: individuals in isolation cannot pro-
vide us with an understanding of the creative nature of the arts because a complete
