I started exploring the idea of “emergence” in a previous entry <here>. Emergence is the idea that something can have a property that doesn’t exist in its individual parts. This could be the fluidity of water or the solidity of ice or even the color of ice. But it could also be the properties of ant colonies that emerge when millions of ants work together and it could be the properties of societies when a large group of humans get together. In fact, in some sense, anything that is not a fundamental principle could be considered an emergent property. And, as I mentioned previously, probably the most interesting proposed example of an emergent property is consciousness. “Consciousness as emergence” is the idea that when a complex network of neurons interact, consciousness emerges from that network even though it is not present in any one neuron.
In the
first entry on this topic, I discussed the emergent property of meaning that
arises from a collection of letters. My
contention was that the emergent property of meaning only exists in one
place: the conscious perception of an
individual. Thus, the “emergence” in
this case can be boiled down to conscious perception and, in the end, the
concept of emergence is exposed to be just another word for consciousness. However, there is some disagreement about
whether the meaning that emerges from letters is really the same as most
commonly accepted examples of emergence, so it is necessary to dig deeper into
this topic.
Today I’d
like to use what I think is a more commonly accepted example of emergence: an image that emerges from a painting. Let’s take the Mona Lisa painting by Leonardo
da Vinci. It is, at a molecular level,
just a bunch of molecules spread over a canvas.
None of the molecules in the paint or even in the canvas have any
resemblance to, or concept of, the image of Mona Lisa that we see when we look
at the picture. Thus, the image that we
see can be said to be an emergent property of the painting or an emergent
property of the molecules of paint that make up the image. The sum of the paint molecules in the Mona
Lisa is certainly greater than just the collection of paint molecules
themselves. There is something more
there – something that emerges when the molecules are arranged carefully (by da
Vinci) that could not have been predicted or expected by just examining the
molecules of paint themselves.
The
question we want to consider is this:
where is the image of the Mona Lisa with respect to the painting? Where does this emergent property exist?
In order
to explain the point I would like to make, I’m going to simplify the situation
with some simplifying assumptions. I
don’t think any of these simplifications affect my conclusion, so please bear
with the simplifications and then we can circle back and the end and consider
whether the real complexity of the situation changes the conclusions I have
drawn.
First, I’m
going to use a printed image of the Mona Lisa instead of the painting
itself. Specifically, a laser-printed
image of the Mona Lisa. Don’t worry –
the image will be immediately recognizable when you look at the page. But, if you were to greatly magnify the
printed page, you would see that it is really just composed of a series of
printed dots of different colors. Each
dot is a single color, based on the ink colors available in the printer. We could go deeper in magnification and talk
about the molecules that make up each dot, but that’s not necessary for the
purposes of this thought experiment. I
think it will be obvious that the image of the Mona Lisa is not resident in the
dots of different colored ink. The dots
themselves have no knowledge of what image they are a part of, obviously. The printer itself was not given information
about Mona Lisa – it was only given information about what color dots to print
where. But there is one thing for
certain: when we look at the page, we
see Mona Lisa.
Now I’m
going to make a couple of major simplifying assumptions. Let us imagine that there is a one-to-one
match between the dots of colored ink on the paper and the photons of different
color being reflected off of the page.
In this case, a single blue photon is reflected off of a single blue dot
on the page and heads to your eye. A
single red photon is reflected off of a single red dot and heads to your
eye. At some given instant in time, all
of the differently colored photons are reflected off of their respective dots
and head to your eye. If all of the
photons were in sync with one another and timed just right, there would be a
mass of colored photons heading toward your eye. Each color and the location of each dot is
preserved in that mass, or column, of photons.
Thus, in some way, you could slice that column of photons and get an
image of the Mona Lisa. But, of course,
the photons, like the dots on the page, have no knowledge of the Mona
Lisa. They are just photons taking a
specific color and existing in a specific point in space. The emergent property of the image of the
Mona Lisa is nowhere to be found in this collection of photons.
This group
of photons now enters your eye where they activate the rod and cone receptors
in your retina. My second simplifying
assumption is that each photon actives a single cone cell, which then generates
an action potential that travels along a single nerve fiber in the optic nerve. Thus, there is a one-to-one correlation
between the whole set of photons and the action potentials on multiple fibers
in the optic nerve. If the nerve fibers
in the optic nerve preserved the geometric relationship of the cones in the
retina, you could slice the nerve at some point and recreate the “digital
image” of the Mona Lisa (after converting the action potentials to their
indicated colors). Again, that is an
oversimplification for the purpose of this analogy. The point is that the light image of the Mona
Lisa lands on the retina and activates the appropriate cones, which send their
signals along the optic nerve to the occipital cortex of the brain. The parallel arrangement of the signals in
the optic nerve are preserved, since each cone cell has a direct “single line”
connection to the brain (due to my simplifications).
The
parallel signals along the optic nerve now enter the occipital cortex, where
they kick off a complex neural network that extends around the visual cortex. Some of one side of the image is split off
and mixes with the other side. Edges are
identified. Colors are identified. The network of neural signals keeps expanding
to different areas of the brain. Memory
is activated and the image is compared to past knowledge and things we
recognize from our past. All of this
happens pretty quickly because it doesn’t take us long to glance at the picture
of Mona Lisa before we recognize it as being Mona Lisa.
Some might
think that eventually these neural signals end up activating some single neuron
in the brain that only lights up when you look at a picture of Mona Lisa. That is sometimes referred to as the
“grandmother cell hypothesis” though I’m not sure that was ever a serious
scientific idea. But there is no such
single neuron. Instead, the collective
activity of the neural networks across the brain give rise to the general
awareness, and thus conscious perception, of Mona Lisa.
My
contention is that, until this final step of consciousness, there is no place
for any emergence of the image of Mona Lisa.
In this whole process, from the dots on the page, through the light
transmission, to the neural activity - all the way to the brain – there is no
place where the information is brought together into a cohesive “image of Mona
Lisa”. It doesn’t happen until it
happens in our brain and we have the conscious recognition of “oh – that’s the
Mona Lisa.”
It is reasonable to wonder whether my simplifications resulted in my “factoring out” emergence from what really happens. Of course the whole process is much more complex. Lots of photons are bouncing off of each dot on the page and heading to the eye, not a single photon. They are coming at different angles, different colors, etc. But photons stay as photons – they don’t mix together to create a cohesive image – so I don’t think anyone would claim that the emergence of the image of Mona Lisa arises from the reflection or travelling of photons from the dots on the page. However, once the photons get to the rods and cones in the retina, the processing is certainly a lot more complex than my simplification. There are ganglion cells in the retina and some pretty complex processing happens even just between the cones/rods and the ganglion cells. My argument here is that the processing in the eye is just an extension of the processing that is carried on in the visual cortex. It can all be lumped under the category of neural processing in the “brain” and, in fact, the retina is considered part of the central nervous system. Thus somewhere, somehow, the image of Mona Lisa arises from the neural processing in our brain (or central nervous system, if you prefer). The emergent property of the image – the Mona Lisa - does not exist until we consciously perceive it. My contention is that every example of emergence (at least every example of “weak emergence”) boils down to this same conclusion: the “emergent property” only emerges in our conscious perception of it, and nowhere else.
~~
I imagine
a discussion with someone about how it is that consciousness just arises from
the neural activity from the brain. I
would maintain that consciousness is non-material [see here], whereas they
would maintain that consciousness is purely a physical thing that naturally
arises from the action of neurons.
“If
consciousness is purely a physical process, what is the means by which it just
arises from neural activity?” I would ask.
“Consciousness
is just an emergent property that naturally arises from the activity of
neurons, just like many other emergent properties of material objects.”
“An
emergent property? What does that
mean??? Could you give me an example
of an emergent property outside of consciousness itself?”
“Well, for
example, imagine you’re looking at the painting of the Mona Lisa…”