Information Technology Reference
In-Depth Information
If so, then turn this back on human creativity. If a creative computer need not
“understand” the value of its own creations, does that mean a human can be deemed
creative even though they are incapable of knowing whether their creations are valu-
able?
To me it seems odd to demand that creativity result in value but allow that the
creator may not know that it does. It would be similar to crediting someone as being
“ethical” even though they cannot discriminate between right and wrong.
My response to these problems is implicit in the chapter I present. I think it will
ultimately be more fruitful to disconnect definitions of creativity from questions of
value.
8
Just as it's a mistake to connect the definition of art to the definition of good
art, I believe it's a mistake to connect the definition of creativity to the definition of
valuable creativity.
I see creativity as being more related to issues around complexity and the be-
haviour of complex systems. For me creativity is simply what complex adaptive
systems “do”, nothing more and nothing less. From this point of view the value of a
given creative act is relative to the (possibly co-evolutionary) situation at hand and
the contribution it makes towards adaptation by the creative entity. In this case hu-
mans, computers, and all manner of things/processes are capable of some degree of
creativity.
PB:
Thanks for this good summary of the situation. It seems to me to hit sev-
eral of the important issues head on. If aesthetic evaluation is uncomputable then
how does the mind/brain do it? As you comment, an interesting question in itself.
As I briefly mentioned previously, it seems to me that the only way beyond this
point is to posit the existence of a metaphysical (super-mechanical) entity which is
unacceptable to me. Therefore I assume it has to be computable.
You infer the work of Gödel and Turing and we know that within any finite axiom
system there will exist propositions that cannot be resolved. However this doesn't
answer the problem since again we must ask: then how does the mind/brain (a finite
system) resolve aesthetic evaluation?
I return also to my earlier mention of Sommerhoff's description of freedom of
will. He implies that things like creativity and aesthetic evaluation may not be com-
putable until the computing engine is at least as complex (or can reflect the same
degree of variety—to use Ross Ashby's term) as the human brain. As suggested in
this discussion, this is a long way off.
Nevertheless we have to start somewhere and it seems to me that starting with
the assumption that computational aesthetic evaluation is not possible is counter
productive—we
must
begin from the belief that it can be achieved.
My glass is half full!
8
This view is also shared by Dorin and Korb in Chap. 13.
