This is the fifth of six parts in my essay on free will.
Perhaps one of the most commonly used decision-making devices today is the computer. Computers make an almost endless variety and number of decisions every second.
Transistors are the fundamental decision-making elements of all computers: the electrical inputs to a transistor determine the electrical output. Combine many transistors, using the outputs of some as the inputs of others, and you can build extremely sophisticated decision-making devices.
Software is essentially a way of controlling transistors in a computer. However, there are many levels of abstraction between the physical foundation of transistors and the set of written instructions making up a software code, as indicated in Figure 4. At the uppermost levels, a programmer might write successful software without even being aware of the existence of transistors. This is not because transistors are unnecessary for executing software codes, but because they are hidden from the programmer by intervening levels of abstraction.
I have two reasons for focusing on computers. The first is that they demonstrate that extremely complex decision-making processes can occur within a natural, causal world. The second is because they are good, if imperfect, analogies to brains. I will discuss each of these ideas in further detail.
First, let me address the idea of complex decision-making in computers. As computers become faster and bigger, it is becoming increasingly difficult for us to distinguish computer decision-making from human decision-making. Computers are able to make complex financial decisions. Given design drawings, they are able to decide which parts are needed to build machines, or what quantities of lumber are needed to build houses. They are able to decide which is the quickest route from one location to another, and they are able to make decisions regarding medical diagnoses. In short, computers are able to make important decisions in just about every area of human endeavor. It is certainly true that computers are not influenced by emotion, but this does not disqualify them from real decision-making. Indeed, emotions are simply additional inputs into the decision-making processes of human minds: they are not essential to the general definition of decision-making.
I should emphasize that computers’ decisions are executed entirely without a supernatural “soul” or “spirit”. All that is needed are transistors and a source of energy to run them. The specific arrangement of those resistors, and how they are activated over time, gives rise to what is known as emergence. Wikipedia describes emergence as “the way complex systems and patterns arise out of a multiplicity of relatively simple interactions.” Emergence is found not only in computers, but in a wide range of physical phenomena (see http://en.wikipedia.org/wiki/Emergence). However, computers have allowed us to get a very close look at the sort of complex behavior that can arise from the interaction of simple physical structures. It seems likely that, some time in the future, we will produce robotic, computer-controlled humanoids whose behavior is essentially indistinguishable from that of humans. When this time comes, we will be faced with very real evidence that personalities, emotions, and complex thoughts can emerge from purely natural, deterministic processes.
This brings me to my second point, namely that the computer serves as a reasonable analog to the brain. Instead of using transistors as their fundamental building blocks, brains use neurons. Neurons are, in principle, a sort of complex transistor: a typical neuron has many inputs, and applies an operation to these inputs in order to produce an output. Neurons even communicate using electrical impulses, as transistors do. And, like transistors, the outputs of one neuron may serve as inputs to many other neurons. Synapses are the junctions at which one neuron’s outputs meet another neuron’s inputs, and the chemical configuration of these junctions also affects the way neurons process their inputs. Brains and computers thus have broadly similar logical hardware making up their lowest level of abstraction. That the brain’s hardware is more complex, with more inter-component connections than a computer’s, simply confirms what we already know: that brains are better than computers at performing certain cognitive tasks, at least for now.
An image of neurons in the neocortex. Credit to Tamily Wiessman (see http://boingboing.net/2010/11/30/mindportraits.html).
Furthermore, like the computer, the brain can be roughly approximated as having different levels of abstraction. Among the lower levels are functions such as basic processing of signals from other parts of the body and from the senses. Higher up are areas of the brain responsible for emotional reflexes like fear, and even higher are areas responsible for cognitive thought. Unsurprisingly, these levels of abstraction appear to have arisen as a result of increasing complexity due to evolutionary processes. First to evolve were primitive nervous systems without brains. An example is the jellyfish nervous system, which serves primarily to coordinate motion needed for swimming. More complex is the central nervous systems of worms, which includes a brain. Insect brains are even more complex, followed by reptile, fish, and bird brains, and finally mammalian brains. As evolution proceeded, brains acquired new, more advanced abilities that usually involved control of pre-existing brain functions.