Digital Digital Burgess Conference Follow-up:
Conference was held August 29-September 1 1997, Banff Alberta, Canada.
I was suggesting that basic physics underlies aesthetics because I'd guess that aesthetics was functional - i.e. it has survival value in some way. I'd suggest that a lot of our aesthetic sensibilities are down to sex, first and foremost. A lot of the geometry of appealing landscapes is reminiscent of the human form (the female form, to my eyes - do women like more rugged landscapes than men?). The place where I live has a lot of ancient mythological associations, most of which involve sex at some deep level, and most of whose landscape features show fairly clearly relevant curve geometry. But landscapes are attractive for other, equally basic reasons - we love "views", and there's certainly survival value in being able to make camp where one can see a long way. Perhaps our sensitivity to landform also allows us to detect and seek out particularly fertile and well resourced locations. This is the sense in which I think Physics is related to aesthetics - aesthetics is a geometric skill (or the side effects of it) related to survival, and maybe the only way to have truly co-evolving "aesthetic agents", as we discussed, is to ground them inside the picture, so that the picture and their survival are related.
But you could certainly skip all that and just develop fitness tests based on human aesthetics, rather than try and get aesthetics to emerge from first principles (the pictures would look better to our eyes anyway!). Given that our aesthetic judgement is, to some extent, reducible to neural phenomena (and I don't mean to demean it by that) - e.g. a selectivity for certain rates of curve, certain ratios and proportions and suchlike, then these could be used as fitness tests. When the Golden Mean was discovered (and I think I mean discovered, rather than invented), that was a breakthrough in our recognition of the components of aesthetic judgement. When a carpenter makes a table that's 5x3, rather than 4x3, he's also tapping into an "aesthetic primitive", even though we don't know what it means. When a photographer places his subject one third of the way into the picture, the same thing applies. Our top-left to bottom-right scanning tendency is another simple cue (and so a photograph of a person and a landscape, when the person is on the left third, becomes a photo of a landscape being looked at by the person, whereas on the right third it becomes a picture of the person standing before a landscape). I wonder whether you could isolate a selection of these "primitives" and use them as fitness tests for an evolved picture? A population of agents could scour all aspects of an image, looking for such primitives - is there a curve with an appropriate expansion ratio? Do any two key dimensions on the picture form a Golden Section?... Perhaps then you'd have a system for evolving art without all the tedium of user selection. Doesn't sound at all easy though - rather you than me!!!!
I'd guess that there are various things afoot at different levels; some of it is fashion-induced abberation (fashions have a dynamic all of their own) and some is perhaps "evolution". The cultural component is certainly strong - nowhere stronger than in the difference between Eastern and Western music, but the underlying cause of that is a mystery to me - nature or nurture?
Our conception of the female form has certainly changed pretty dramatically since the Stone Age, judging by carvings! Yet perhaps the rule has remained constant but the circumstances changed - perhaps aesthetics are still based on survival value, but the criteria for survivability have changed. Maybe in the Stone Age large, well-rounded women had a greater chance of bearing healthy children, but nowadays some other sexual selection criterion has come to the fore that makes supermodels have higher survival value? Certainly they tend to earn more than fat women, but that's a circular argument! Either way, the fashion for thin=attractive is too new to accept genetic evolution as the cause.
Like everything else, I expect what we call aesthetics is actually many interrelated phenomena. Perhaps there are certain constants, such as the ones Steven suggested about light and dark rectangles - these are obviously directly related to natural phenomena and might even be hard-wired into our visual cortex. Then perhaps there are more sophisticated and less hard-coded ones that are more susceptible to the "attractors" of fashion and culture. Both are examples of a visual feature correlated with a "goodness" value, but some are learned and some innate.
Darrel's comments about the relationship between aesthetics and new media are certainly food for thought, but this is getting way beyond my field and I'm rambling, so I'd better shut up! Is there a psychologist in the house?