Relative to the previous vehicles, this chapter explores two additional kinds of design innovations. The first is inhibition -- the motor output now becomes inversely proportional to the sensor output. The second is the use of multiple types of connections, including inhibitory and excitatory, crossed and uncrossed, and a variety of different types of sensors.
If we take vehicles 2a and 2b, and make the connections from the sensors to the motors inhibitory, the motors will slow down as the detected signal strength increases. Furthermore, the motors will really race when the signal is weak. The two systems will come to rest in the immediate vicinity of the signal source.
The system with uncrossed connections will, in this case, come to rest as it turns directly towards the source, because (if the source is to one side), the motor nearer the side of the source will slow down more than the motor on the far side. In Braitenberg's words, this vehicle "LOVES [the source] in a permanent way, stying close by in quiet admiration from the time it spots the source to all future time."
In contrast, the vehicle with crossed (inhibitory) connections will come to rest facing away from the signal source. One consequence of this is that the effect of the source on the sensors is weakened (because they point away). If this weakening is sufficient, the motors could start up enough to drive the vehicle away from the source. This vehicle "is an EXPLORER. It likes the nearby source all right, but keeps an eye open for other, perhaps stronger sources, which it will sail to, given a chance, in order to find a more permanent and gratifying appeasement."
Braitenberg now proceeds to design Vehicle 3c, which builds upon all of the principles seen so far. This vehicle has four sets of sensors, each tuned to a qualitatively different kind of signal. The first sensor type is sensitive to light, and has uncrossed excitatory connections to the two motors. The second sensor type detects temperature, and has crossed excitatory connections to the two motors. The third sensor type detects oxygen concentration, and has uncrossed inhibitory connections to the two motors. The fourth sensor type is sensitive to amount of organic material, and has crossed inhibitory connections to the two motors.
"This is now a vehicle with really interesting behavior. It dislikes high temperature, turns away from hot places, and at the same time seems to dislike light bulbs with even greater passion, since it turns toward them and destroys them. On the other had it definitely seems to prefer a well-oxygenated environment and one containing many organic molecules, since it spends much of its time in such places. But it is in the habit of moving elsewhere when he supply of either organic matter or (especially) oxygen is low." Braitenberg describes this machine as having VALUES and KNOWLEDGE.
To Braitenberg, this thought experiment immediately raises the issue of how knowledge is defined. "But, you will say, this is ridiculous: knowledge implies a flow of information from the environment into a living being or at least into something like a living being. There was no such transmission of information here. We were just playing with sensors, motors, and connections: the properties that happened to emerge may look like knowledge but really are not."
In other words, the definition of knowledge differs from the behavioral perspective (watching the vehicle) vs. the architectural perspective (knowing how the machine was built).
To my mind, this thought experiment very nicely brings to mind a central notion in cognitive theory -- at some point, "magic" (knowledge?) must be eliminated from a functional theory if it is to be viewed as a scientific explanation. In short, the conflict between the behavioral and architectural views of knowledge is exactly the type of conflict that cognitivists should be striving to create!