In an earlier post, I wrote:
That leaves, as one of the basic problems for a cognitive agent, the problem of getting information about the world.
In this post, I want to discuss why that is a problem.
Many people seem to hold the view that sensory cells in the body passively receive input from the world, and that how we perceive the world depends on what we do with that passively received data. That seems to be the view of proponents of sense-data accounts and of proponents of computationalism.
The traffic light controller
I’ll begin by discussing the traffic light controller, which does passively receive input. Then I will contrast that with the situation of a cognitive agent, say a policeman directing traffic.
A typical traffic light controller is a small computer in a weatherproof box, which is able to switch the traffic lights at an intersection. There is often a detector for each lane of traffic entering the intersection. The detector might be installed under the roadway, and is connected to the traffic light controller. The controller might have additional inputs, perhaps allowing it synchronize its activity with other nearby traffic lights. For the purpose of this discussion, I will mainly concentrate on the traffic detectors for each lane of traffic.
The controller does not have to do anything to receive the inputs from those traffic detectors. That’s why we can consider it a passive receiver of inputs. But, once the inputs are received, the controller uses those in its computation to make decisions on when to switch the lights to allow traffic from a different direction.
We see the inputs to the traffic controller as meaningful, because they are wired to the lane traffic sensors. The meaninfulness of the inputs comes from the causal connections to the lanes sensors, and by virtue of the causal connection between the sensors and the traffic passing over the sensors. It is much disputed as to whether the computer in the traffic controller is actually capable of having meaning. However, in our talk about how the traffic controller works, we often attribute meaning to the computer. It is a common view among computationalists, that there is nothing more to meaning than attribution. That is roughly the view argued by Dennett as “the intentional stance.”
The traffic policeman
Now suppose that there was a traffic accident at the intersection, and that some of the lanes are blocked by the damaged vehicles. A traffic policeman is called in to direct traffic, because the traffic light controller is not programmed to handle the rerouting of traffic to unusual lanes.
We can consider one of the sensors of that policeman, say a blue sensitive retinal cell. One moment it might be stimulated by a photon from the sky. A moment later, it might receive a photon reflected off a blue automobile. And, an instant later, it might receive a photon from the sky that was reflected of the windshield of another automobile.
There is no consistency to the source of the stimulation. So if we look at passively received signals, then what is received is about what William James described a “a bloomin’ buzzin’ confusion”. There is no meaning that can be determined from which sensor was stimulated, because there is no clear causal connection to the world. Trees and rocks have it much easier, because at least they have a fixed orientation, so that there is some sort of causal connection.
We might think of the possibility of the brain taking such passively received data, and adjusting it according to the orientation of the policeman. But there’s a difficulty here. In order to receive reliable real world information, the brain would have to first know the orientation. And in order to determine the orientation, the brain would first have to be capable of receiving reliable real world information.
There are possible ways out of this dilemma. But those ways depend on the brain actively seeking real world information, instead of just passively receiving it. I’ll discuss that in a future post.