This is a post about a painting that I completed only last week. But it was a painting that I’d been trying to figure out how to paint for all of 7 years. It all started with a motorbike ride.
One day in 2006, I was riding down a mountainous road in North Carolina,
when I saw a hill of striking beauty (I believe it was somewhere near a town called
It was evening and the colours were beginning to fade, but I could clearly
make out on one side of the hill striking trees that were reddish-orange in hue and I
found them beautiful against the light green that surrounded them and the blue skies
I continued to ride down the road. I rounded the hill and came to the other
side. I saw again, waves of green and yellow and touches of red and I was overcome
with a sense of being in the presence of the highest beauty that the eye could
So, I stopped and removed the camera from the knapsack behind me. From
various points along the road, I tried to photograph the hill.
However, I could not frame a shot that could evoke the beauty of
the hill that stood before my eyes.
That evening, I was troubled by a sense of failure, of bafflement at having failed to record the soul of the scenery that I had left behind. The hills around Marion were all extremely beautiful, but nothing else on the road had the perfection of the hill that I had not been able to photograph.
So, that evening, I sat and pondered a very interesting question that had to do with
aesthetics. The question was this: Does the mind experience beauty according to
images that it has seen, or does the mind experience beauty by a mental model it
has built of an object and the associations it has made between that model and
models of other objects?
Let me clarify what I mean by a mental model with an example. When we approach a
person whom we have never met before, we don’t immediately see all of that person.
We perhaps at first only see the front of the person, at a distance. Later we see more
of the person, their face, their profile, and how they look from behind. As we see
more and more of a person, we fill in more and more of the missing information about
how the person looks.
However, from the very beginning, we know that we are
looking at a person, even though we haven’t seen the person’s teeth or sides or
back. So we must have decided that the frontal image that we saw fit best into a
mental model of a person. The mental representation may not just be a 3
dimensional representation. It may be a set of associations and memories over time, all woven together in complex ways.
So, the question that came to my mind was: Is a human mind’s evaluation of
the aesthetics of an object based on a) any of the individual images of the object
that we have seen or on b) the entire mental model that we have constructed in our
minds of the object.
The more I thought about it, the more I began to feel that my experience with the hill
would be best explained by the conclusion that our aesthetic experience of an
object is determined by the mental model we have formed of it, and not by any image
of it, and here were the reasons for my leaning towards the latter conjecture:
a) If it is just a single view or image of an object that drives us to experience it as
beautiful or attractive, then our feelings about its attractiveness should change all
the time, based on small things like the angle at which we are viewing it. That does
not seem to be how we experience things in real life.
b) The experience of the hill led me to think that the latter explanation was the more
likely because if it was one of the views of the hill that made me think it beautiful, I
should have been able to capture the view with my camera. Since I couldn’t,
perhaps it was something I saw on the other side of the hill, combined with the
fading light and the landscape along the way (perhaps combined with the beauty of
some of the trees I saw upon the hill) that came together to contribute to the sense
of beauty that I experienced when looking at the hill.
That evening, I thought of a new kind of painting technique that would be possible if
the above conjecture proved true. It seemed that it might be worthwhile to paint a
picture from the point of view of the mental model a viewer might form if they saw the actual subject of the painting in context, possibly from different angles, or at different distances, and over an extended period of time.
But it was only last week, a good 7 years after I passed through Marion, NC, that I
tried to create a painting of the sort. Since I have recently watched many Indian
classical dance performances and some contemporary dance performances, I
attempted to paint all the things that might be associated with a dance performance
in the mind of a spectator. I constructed the painting like a fractal, starting with a
dancer and then representing the dancer at different levels of granularity and detail.
One of the things on the mind of the spectator would be the dancer’s body
(the whole form, from head to toe). Emoting is a key aspect of dancing in Indian classical dance. So, in order to appreciate the emotion, a spectator would have to pay close attention to (and so form a model of) the dancer’s face, and of his or her eyes. So, I wove an
image of a forehead, and of an eye, into the painting. Another part of the body that
Indian classical dancers use a lot is the hands and the fingers. They use them to
denote various objects, characters and actions. I associated the movements of the
arms and fingers with flowing water, and plants with branches. So, I painted in some
flowing washes and let some negative white spaces take the form of trees in the
painting. Indian dance costumes are colourful, so I made everything in the painting
as bright as a classical dance costume.
The process of painting was a lot of fun. It was very different from the process of
scientific experimentation of course. There was none of the same rigour and I
wasn’t really looking to prove anything.
I liked the painting that came about and I have inserted a photograph of the same into this blog entry (it’s at the top).
But we don’t know yet whether mental models lie at the root of aesthetic
experiences. And you know what? As an AI researcher, I am just dying to find out! It
seems possible to design an experiment to determine the validity of the conjecture.
We could take an object whose facade is beautiful but whose side view is not. The
subjects of the experiment would be shown the front and side views of the building
and asked to rate the appearance of the front of the building. The control group
would be shown just the front of the building and asked to rate its appearance.
A significant difference in ratings (poorer ratings by subjects shown the front and
side view) would support the conjecture that a mental model lay at the root of the
I hope that one fine day, we might be able to conduct the experiment, and if we do, I
shall share the results here with you!
As a footnote, I’d like to share this blog post http://aimeeknight.com/2012/04/10/the-new-aestheticperhaps/ by a blogger from (it appears to be – hmm, what a coincidence) North Carolina. The article is about art that attempts to portray the world from the point of view of a machine – using a machine’s aesthetic – the phrase they’ve used is ‘New Aesthetic’. I had not heard of the concept before, and it isn’t the same thing that I am working on, but since we’re discussing aesthetics and perception, it appeared relevant and interesting.
As a second footnote, I’d like to put in a word of explanation about why a firm of computer scientists and AI researchers should be interesting in painting. Well, to many early AI researchers, AI was as much a quest to understand the human mind as it was a quest to solve difficult problems. There were many attempts to comprehend the human mind by constructing logical and mathematical models of various cognitive processes (things the mind did) and seeing if the models could mimic the real thing. So, AI research had a lot of common ground with Psychology. And there is still a cross-disciplinary research area known as Cognitive Science that blends Psychology and AI research. So, I just want to defend myself by saying that talking about aesthetics is not off-topic for this blog.