[Previous chapter]

Chapter 4: All The Answers Revealed- I Think

Now we have some heavy philosophical artillery under our belts. We are ready to tackle the Questions, as well as the cynicism of Chapter 2. In this chapter I will describe a fourth theory of what ethics is: correct-ism (because I believe it to be correct!) It is similar to boo-yay-ism, but it includes a lot of refinements. I strongly believe that correct-ism is true. However, do not forget that many very smart people have been arguing about these issues for thousands of years. I hope that you either find this chapter satisfies your confusion, or else I hope you will tell me what you disagree with.

And the Winner Is…

We need to make several decisions. We have learned two stories about how the mind makes decisions: the One-Driver Story and the Many-Driver Story. Either we use Reason to decide amongst our desires, or else it is the strongest desire that always acts. We have also learned three different theories of what ethics is. Happy-ism says that "good" and "bad" are just new words for another, non-moral property. Sixth-sense-ism says we have an intuitive sense for the value of things. And lastly, boo-yay-ism suggests that when we call something good, we are saying that we are motivated to pursue it. We need to pick one of the two stories of decision-making, and one of the three theories of what morals are. How will we decide?

The first important fact to consider is that people are machines, even if it does not feel like I am one. If there is one thing we can learn from science, it is that everything in this world can be seen as simple parts interacting to produce more complicated behavior. From medicine, we can see that the human body is made of parts, that each contributes some function to our survival. The alternative view is that there are "supernatural" things in the world, which are complicated but cannot be broken into parts. People who believe in a spirit that controls the rain would hold this view, for example. We have found through experimenting that we can always break up a system into smaller parts, until we get to the smallest subatomic particles. I think this assumption is a valid one to make, because without it much of our technology would be impossible. If there were no underlying simple regularity in things, we could not get repeatable results with experimentation. And since science has shown itself to work very well – which we can see any time we microwave popcorn or use a computer – we have good reason to believe that science is on to something.

So if people are machines, we must get rid of the One-Driver Story. The problem with this story is that it makes Reason to be something different from a machine, rather a pure distillation of Will. The behavior of a machine can always be predicted in principle by studying the behavior of its parts, but the One-Driver Story makes Reason to be totally free. It is not predictable, because if you could predict everything I did I would feel controlled somehow, not my usual free self.

The One-Driver Story tries to avoid an important question: how do our brains make decisions? It feels like sometimes we are free to choose, but other times our minds get the best of us. Sometimes we eat too much candy before dinner, even through we know we’ll feel sick. So it is not the whole brain that makes the decision. If there is One Driver, then it is just the little Reason part of the brain that makes the decision, and the other parts of the brain sometimes get the best of it. But if Reason is how we make decisions, then how does Reason make decisions? Since people are machines, and minds are the operation of brains, Reason must be a machine built out of neurons. The One-Driver Story says that Reason is supernatural – a complicated force in the universe that cannot be understood in terms of simpler parts. We should be able to explain how it works, but the One-Driver Story just avoids the question.

The Many-Driver Story, however, solves this problem. It begins by showing one way to divide the human machine into parts. Our decision making can be divided into several desires, each of which is a neural circuit that leads our body to behave a certain way. One circuit wins, and gets to execute. Although our behavior is predictable in principle, however, we do not need to worry. Our brains are so complicated, that mind reading may never be practical, certainly not for many hundreds of years.

So how do we explain feeling like our mind got the best of us, as when we eat too much candy before dinner? Well, here we desire two things. We want candy, and we want to not feel sick. Reason fills in some details: "If you eat candy now you will feel sick later." The desire to eat candy is stronger, so we take some. However, the desire for food is usually short-lived. A moment later, we have no desire for candy, only the desire to wait until dinner. With this new set of motivations, our previous decision seems incomprehensible. "Why did I do that?" we ask ourselves. We made a bad decision because we followed a short-term desire rather than a long-term one. The long-term one sticks around to haunt us later, and we feel regret. Hmmm…. This is starting to sound like moral decision-making, eh? But before we get ahead of ourselves, let’s go back to the theories of what morals are.

One Decision Down, One to Go

If people are machines, then sixth-sense-ism cannot be right. This theory says there is another property in the world that is very complex, the property of how good or bad things are. It cannot be broken down into parts at all. Rather, we know about it by feeling it directly. However, it seems very unlikely that this is true. If the world is a machine made of subatomic particles, value would be a totally different aspect of the universe than anything else scientists know about. There is no way to measure value, as there is to measure speed or mass. And where is the value-sensing organ of the body, as the eyes detect light? Sixth-sense-ism, even though it sounds like how we feel, cannot be right.

So we are left with happy-ism and boo-yay-ism. I believe that boo-yay-ism must be right, because of the motivation problem. If happy-ism is right, why am I motivated to be good? I am not always good, of course, but I sometimes am. Happy-ism says that "good" and "bad" are just new words for some other property of things in the world. But no other property has this power to make me act. As I mentioned earlier, I do not automatically want to do whichever action increases the number of pigs on the planet the most, or whichever action helps my next door neighbor learn to play the trombone. Happy-ism fails to explain why morals are different, so it cannot be right.

The winner, then, is boo-yay-ism. Calling something "good" means that I feel motivated to do it. But boo-yay-ism has important problems as well. First, I am motivated to do many bad things as well as good things. Secondly, boo-yay-ism cannot explain why morals seem to be objective. Let us look at how each of these problems can be solved.

So we cannot simply say that "good" is exactly the same as "whatever I am motivated to do." We could not answer this objection before, but we now have the tools with which to do so. Think back to several pages ago. I discussed how the Many-Driver Story can explain "our minds getting the best of us." Our desires change over time. We can plan to pursue one thing, only to have some new desire burst onto the scene and wreck everything. When the sudden desire goes away, we are left regretting how our plan has failed. The thing we desired for a moment was not good. The good motivation was the one that stuck with us afterwards. So what separates "good motivations" from "bad motivations" is how long they last. We never want to feel regret, and yet we know that our motivations are always changing. We plan ahead with this information, doing "the right thing" even when "we don’t want to." By always keeping in mind what is good, we will not be faced with regret. The motivations that are always in the background are being kind and giving, acting selflessly, and other drives that we call "good." The motivations that come and go are the "human weaknesses" that often spoil the plans of the quieter ones: hunger, lust, pride, etc. So "good" is not simply "whatever we feel like at a given moment." Rather, it is what we want all the time and what we will want a few moments later – the desires that stick with us.

So now we can move on to the second problem with boo-yay-ism. Why do morals seem somewhat objective? Why do I need to observe the same moral rules as everyone else? To answer this objection to boo-yay-ism, I’m going to tell a little story…

The Self-Improving Robot

Imagine a robot that is designed to improve itself. It has the ability to upgrade both its hardware and software, with the goal of making itself able to compute as fast as possible. The robot has access to a computer chip manufacturing facility, equipment for fabricating circuit boards, and other similar resources that allow it to create new hardware. It has access to a computerized physics, computer science, and engineering library in order to "learn" new things. It also can perform scientific experiments in order to discover new things that it has not already been programmed to know. The human designers of this robot give it some initial software to begin its task. When it is first turned on, the robot is programmed with a few simple ways to use its resources to increase its speed. As it develops, however, the robot can change its own program to "think" in new ways, or to compute things with algorithms it did not previously have.

This robot is not possible with today’s technology. We are not capable of writing the initial program, which would have to allow reading of the digital library, many kinds of physical interaction with the outside world, and other computationally difficult tasks. If we don’t let these technical problems stop our imaginations, however, think of what would happen. The robot is a machine, and can only follow the rules with which it was programmed. It begins with a certain hardware and software design. Using one design, however, it creates a better design, which can in turn create a still better one. Each design is faster than the one before, and therefore capable of tackling harder design problems. With faster speed, the robot can run new algorithms to invent novel designs. As time goes on, the robot evolves, and grows faster. Suppose that several such robots were activated in different places at the same time. Each was given slightly different materials to work with, or slightly different knowledge in the original program. What will the situation be after the robots have had several years to work?

It is quite possible that each will look quite unique. Each robot may follow its own path forward. One knows more about circuit engineering, and so begins with a hardware improvement. Another begins with more knowledge of software design, and recognizes an inefficiency in its program. One robot might be "lucky" enough to discover a radical new design, and scrap most of its existing parts. At the same time, however, there are certain general features we would expect to be shared by all robots. All will develop smaller, faster computer chips. All will strike a balance between spending time on experimentation, scanning the scientific literature, and executing design algorithms, all sub-goals that assist the main goal of increasing speed.

Is there a "best" design? It is hard imagine that we are anywhere close to the perfect design when we begin. However, the robots do get faster. There are physical limits to how fast information can travel, so there is some limit to how fast the robots can become. At some point in the distant future, the robots will not be able to make any improvements. They are constantly pressing onward, getting closer and closer to that goal.

Now what could any of this have to do with morality?

The Self-Improving Person

You probably could see where I was going with the robot story. Design changes are objectively good or bad for the robot in the same way that choices are objectively good or bad for people. The robot is a machine, executing rules. But the rules lead it somewhere. There is such a thing as a better robot- a faster one. In the same way, people are machines whose brains apply certain rules to produce behavior. But that behavior is still leading somewhere. So even if boo-yay-ism is correct and morality is defined by what we want, we can still have objective right and wrong. Before we had trouble with boo-yay-ism because it seemed to make morality too subjective, but in the robot example we can see how it can actually be objective.

How can a robot be objectively wrong? Suppose the robot Speedy Sam shares part of its design with another robot, Dumb Dan: "Dan, you could work even faster if you powered yourself with five volts rather than 120 volts." Dan has rules for figuring out which sorts of designs will be fast or not. But his program is not perfect. Dan thinks that Sam’s design will actually be slower. "Sam, I need to run my big mechanical arm. Five volts would not be strong enough to make my arm work, and without my arm I cannot move anything." With the design that Dan currently has, Sam’s suggestion seems bad. But Dan may still be wrong. By using a smaller power supply, we can imagine that maybe Sam generates less heat and can thus use superconducting circuitry. [6]  If Dan knew everything Sam did, he would see that the suggestion is really a good idea.

Now let us switch back to people. In this example Speedy Susie tells Dumb Denise, "Denise, you should not use slave labor to work your fields." Denise might believe otherwise saying, "Without the work I give them, my slaves would not have the benefits of civilization." To Denise, this really does make sense, given her beliefs at the moment. But Denise is objectively mistaken. Perhaps her desire for large profits is making her reasoning shaky. She does not realize that she is actually doing her slaves more harm than good. If she cleared up this error in reasoning, she would then see why Susie’s suggestion makes sense.

Just as the robots are all programmed to pursue the same goal – processing speed – people are programmed by evolution to want the same things at heart. We want food, shelter, companionship, and we want to survive. As the robots go down different paths to pursue their goal, we develop different cultures to approach our goals. In both cases, this means that there are many differences between individuals. Two robots or people might very well be unable to resolve any particular difference at once. What seems good to one person might seem dumb to another, because of the other differences lurking in the background. If we resolved those differences, however, we would end up agreeing. As one robot really is faster than another, one person might really be more effective at achieving the basic human needs than another. The boo-yay-ist can still be objective about morality.

So to Sum Up…

So here’s what we’ve found. People are machines, with many general rules for behavior (like "eat!") running around in our brains. Some of these rules, but not all, are carried out and become behavior. Those rules that stay pretty much the same are the basis for morality. We have some behavior rules that randomly seize control, but we have a reason to suppress them because we fear a future sense of regret. With the ground rules for morality essentially the same across all people, cultures come up with different specific rules of behavior (like "go to the grocery store") to achieve the basic ones. Some cultures really are better at getting what we basically want, so they are objectively better than those that do not.

Whew!

So What Are the Answers to the Questions?

So let’s summarize: Why should we be good in general? Because we will regret it otherwise. Why should we do good things like help the poor and be kind to strangers? Because our life is better in a world where we do such things.

[6] I am speaking hypothetically, of course. I have no idea what power source the fastest possible robot would use.

[Next chapter]

[Return to writing page]