In the last section I argued that if an agent is governed by necessity in the sense that there is a real inability to do anything else, that agent has no choice in the matter. I also argued that the real inability must be in the mind if it is to truly eliminate choice, for choice is an attribute of the mind. In this section I present the attributes of choice.
1 The Concept of Freedom
Freedom, as I will present it, is the exact opposite of necessity. Freedom implies choice in the sense that an agent that is free may be properly blamed or praised. Freedom does not mean that rules are absent, but that the rules are of obligation and not necessity. It is precisely this nature of the rules of freedom that make true blame or praise possible. Linked closely with these concepts are the notions of punishment and reward.
Freedom and necessity are opposites and must not be confused. It is nonsense to speak of persuading an agent of necessity or forcing an agent of freedom. Agents of necessity cannot be persuaded to do anything other than what they do. They do what they do simply because they can do nothing else. No amount of persuasion changes this nature. Similarly it is impossible to force an agent of freedom to choose any particular option because if the choice is by force it is not a choice in the proper sense and the agent becomes one of necessity rather than freedom.
In the preceding paragraph agents of necessity are described as doing, and agents of freedom are described as choosing. The reason for this distinction is clear from the nature of the case. Agents of necessity behave according to their context and there is nothing in back of this behavior that is not determined by the context. On the other hand, agents of necessity make choices that are not necessarily consistent with their outward behavior. For example, I may choose to lift my hand above my head but in the process of doing so someone grabs my hand and holds it down. The fact that I did not do what I chose does not necessarily mean that I did not choose to do it. Therefore agents of necessity just do and cannot be judged for what they do in the sense of assigning praise or blame. Agents of freedom make choices and can be judged for their choices for the purpose of assigning praise or blame.
2 Choice is Free From Constraints
A common view of choice as argued in blog 2 is that choice is a system of constraints. In other words, presented with a number of options an agent collects constraints until only one option is available and that is the one chosen. In this view constraints eliminate options by necessity in a manner similar to how an optimizing program works in a computer. For example, if a program were designed to find the minimum of the function y = x2 and the present value of x is 2 the program must determine which way to change x to get closer to the minimum. It should either increase x or decrease x. Does the program have a choice in what it does? Absolutely not. If programmed properly it will decrease x because y(2.1) is bigger than y(1.9). This mathematical fact or constraint determines the direction of change for x and there are no options. There never were any options because the context completely determined the response. The program simply maps the context into the response which was already, in essence, contained within the context.
Choice is free from constraints in that to be choice at all, it must not be determined by necessity. Therefore if an agent of freedom were solving the minimization problem described above it might determine the same facts about y(2.1) and y(1.9) but would be completely free to choose increasing x despite the evidence that x should be decreased. A computer program is not free to make such a choice.
Another example that illustrates the principle is the following. Suppose a coin that may be biased is tossed and you have the choice of calling heads or tails. Before the first toss you have no idea what to choose so either heads or tails seem reasonable. If the coin is tossed ten times and comes up heads every time, then it seems reasonable to choose heads. But you are still perfectly free to choose tails. The fact that ten of the ten first tosses came up heads does not create within you a real inability to choose tails.
Now suppose a computer program was prepared to take evidence from previous tosses into account. It might do this by determining the side that occurred the most often and use that as the prediction. In this situation the prediction is entirely determined by the context, i.e., the previous tosses. The program makes no choices because it is not free to do but what it does. It is the programmer who had the choice in the sense that he could have chosen a different rule for making the prediction.
Therefore we conclude that evidence does not determine (in the sense of making it necessary) choice.
3 Choice Implies Ability
In discussing choice as an offering in blog 2 it was determined that inability to select an offering nullified the existence of choice. Therefore the word choice implies that the agent has a real ability to choose, or not choose, a particular option. To illustrate this concept in yet another way, consider this thought in the context of commands given to an agent. Commands are usually given with the intent that the agents to whom they are addressed obey them. Commands given to agents of necessity are quite different from commands given to agents of freedom in the sense that the whole framework for the command is different. Commands given to a robot, for example, simply change the context of the robot such that it must follow the command. The robot, being an agent of necessity, simply reacts to the new context and responds accordingly. Viewed externally the behavior of the robot might not be consistent with the command but nevertheless the robot does what it is actually commanded to do regardless of what the command means to an external observer. For example, I might program a robot to turn left whenever the word right appeared in a certain memory location. An external observer might think it strange that the robot would turn left at the command right, but it would make perfect sense to me. On the other hand I might simply have made a mistake in programming and it may be a surprise to me as well, but the action of turning left is still by necessity. The robot has no real ability to do anything else, hence it has no choice.
An agent of freedom, given the command to turn right, might turn left or right or maybe not at all, depending on the options available and the choice made. While the robot could not be praised or blamed for what it did, such a possibility does exist for the agent of freedom, but only so far as a real ability exists. For example, if the agent of freedom were commanded to go right and it decided to follow the command and go right, but a force stronger than the agent compelled it to go left, the agent of freedom could not be blamed for its action. The choice of the agent of freedom is in line with the command but there is a real inability to go right. Similarly, if the agent of freedom chose to go left, in violation of the command, but was compelled by a force stronger than it to go right, praise cannot be properly attached to the action.
Continuing with the theme that a proper ability must be present for choice to exist, consider the above example at the level of choice rather than action. Does any principle change? No. If somehow the options of the agent of freedom are reduced to one by a constraining force beyond the agent’s ability to resist, there is no real ability, no choice, and the agent becomes one of necessity, and is neither praiseworthy nor blameworthy.
In conclusion, for choice to exist there has to be a real ability for the agent to select the different options presented. To speak of choice without a real ability is to speak nonsense.
4 Choice is Reasoned
In blog 2 choice as randomness was discussed and it was determined that true randomness possessed options and a real ability to select the options and therefore has two of the necessary properties of choice. However, the concept of a random process being praiseworthy or blameworthy was a little hard to swallow. So what is this mysterious third ingredient that makes choice more than a random process? Choice is reasoned. This does not mean that reasons act as constraints because then there would be no choice. The concept is more like this, a choice is made for a reason while at every moment the choice could have been otherwise. In other words the reason does not dictate the choice by necessity, but is considered in making the choice. Is there a mystery here? Yes, I would say that there is. But this is precisely where the mystery is and nowhere else. It appears to me that our common definition of choice which encompasses the concepts of praise and blame developed as a description of this mystery. It does not matter that we do not fully grasp the how, as long as we continue to understand the principle. Explaining choice as a possible attribute of an agent of necessity and blatantly ignoring the need for a real ability to select options does not add to the understanding of choice but simply changes the definition to fit something that is understood.
An example may help make the point. Suppose a person is asked to choose a red ball or a blue ball and told that one of the balls has a gold bar hidden inside. They are allowed to x-ray the balls and measure their specific gravity etc. to provide as much evidence as possible to verify the presence of the gold bar. Now, if all the evidence confirms the presence of the gold bar in the red ball, is the person still free to choose the blue ball? Yes. Even if the person is shown the gold bar and desires to choose the red ball they are still free to choose the blue ball. The evidence and even the desire do not create a real inability.
Explaining choice in agents of freedom as if they were agents of necessity serves no purpose in furthering the understanding of choice. Every time we issue a command, ask for a commitment, bestow praise or impute blame we deny that choice is an attribute of an agent of necessity so why pretend that it is?
0 Comments