In <m0s4knS-0007GNC@rbdc.rbdc.com> Brandon Van every wrote:
>
>
> Confessions of a bit-head...
> I am not as familiar with Dali as I'd like to be.
>
> The best way to change that, would be to go to your local library and
> get a book of his work.
>
> I know this may be a bit painful to some of you art majors ;-), but how
> would you describe Dali and his interpretation of things, if you were
> going to build an algorythm for "Dali-izing" say, a Buick?
^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
> Auto-sodomize it with a melting skull. Then fill it with cauliflower.
> I am dead serious, this is what you'd do to "Dali-ize" it.
Slightly off-topic and in a more general response to the notion of
modelling creativity, Charles Tijus (1988) suggests the solution that,
"Generally speaking, scientific expertise is based essentially on deductive
inference, whereas the processes of artistic creation use inductive inference
as well."
Tijus sees invention as the root of creativity, concerning,
"...a new combination of elements that could not have been discovered earlier."
His conclusions are based on research using the Critical Incidents Technique
by J C Flanagan (1954) by which method data was collected on the creation
processes of 31 artists.
"This technique permits one to reconstitue retroactively the thoughts that
led to the realization of an artwork judged as innovative by the artist."
He describes a machine model of an "Association Discovery Model".
"Long-term memory is represented frequently as associative chains in a
semantic network which are symbolized by a tree graph with nodes and links,
for example; "house" could be linked to "kitchen" and to "bedroom" as well
as to "brick", with an argument for each link."
The weight of each argument is updated by the network in the light of new
information based on repetition and performance. The second level of the
network uses "hypergraph" and "spaces" allowing the notion of "domains",
the concepts are considered as belonging to a particular domain such as the
"home's domain". The third level is concerned with the interpretation of
texts or tasks.
"More often than not, in the history of art, the artwork involves an
association between things, between universes that originally did not belong
together."
Tijus proposes for example, that when Picasso painted "Les Demoiselles
d'Avignon" he made an association between African art and contemporary
painting and that for the Surrealists it was the union of dreams with
contemporary representational painting; for Cesar the auotomobile with
sculpture and with Moore the "hole with the full".
"These associations are more or less strong (in the sense that they are more
or less unusual) according to the semantic distance that separates the two
universes bridged by the created product."
He submits that there are five procedures that contribute to the production
of new ideas in artistic practice.
(1) Focused and Continuous Innovation
(2) Focused and Sudden Innovation
(3) Sought-After Innovation
(4) Natural Innovation
(5) Problem-Solving Innovation
So the wider the semantic distance between the elements in a composition
the more surreal it appears?
> Then fill it with cauliflower.
I like that a lot :)
Regards,
Steve
COGNITIVE PROCESSES IN ARTISTIC CREATION:
TOWARDS THE REALISATION OF A CREATIVE MACHINE
Charles Albert Tijus
"Leonardo", Vol 21, No 4, pp 445-452, 1988
-- __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ scockrill@sprill.demon.co.uk Steven Cockrill - London