An online component of a doctoral study, Theory and Practice of Handmade Distributed Representation, based at the University of Leeds (UK) and funded (2001–2004) by the Arts and Humanities Research Council (AHRC). The full text of the PhD thesis is available on the site.
The thesis also attempts a detailed account of the motivation for the development of ASNakedGene: to find a way of addressing experimentally one of the most urgent, yet oldest, concerns in art practice, known in the folklore (e.g.) as the distinction between being ‘in control’ and ‘letting go’, but also variously scrutinised via the related distinctions between: intent and intentional arc (Merleau-Ponty), sacrificial and transcendental blindness (Derrida), form and formless (Krauss), centripetal and centrifugal forces (Bakhtin), the differentiated and the undifferentiated (Ehrenzweig), attenuation and repleteness (Goodman), pragmatic/rhetoric and epistemic investments (cf. Kirsh and Maglio), representation and distributed representation (in connectionism), etc. In hope of unifying these perspectives (and following Dennett’s elaboration of Simon and Valéry), the problem of the articulated generator is identified as central to the study of creativity. Furthermore, the most concrete expression of the articulated generator was found in the problem of genetic representation in (creative) evolutionary systems, and specifically in evolution by “aesthetic selection” (Dawkins). In an effort to establish a germane evaluative framework, the concept of lossy incompressibility is introduced as the relevant notion of complexity (in contrast to Kolmogorov complexity, which equates it with randomness, i.e. lossless incompressibility).
ASNakedGene itself was a novel implementation of evolution by aesthetic selection of images where (1) there is no genotype to phenotype mapping, no germ-line/somatic-line distinction, no genetic representation, and where (2) evolutionary runs can be seeded by arbitrary sets of digital images.