My View on Free Will
Okay, I decided that I will write my (somehow unique) view of free will in plain English. I have told this view point to several friends of mine and successfully convinced them, but this is the first time I put this in English (so that my friends that are born and raised in the country of “freedom” can also see my view.) As a second language speaker, I am not a very good English writer, so please bear my writing (or at least bear my not being confident).
The world is not deterministic, as physics have shown. This means that we can say that, at any point of time, there are many possible futures. However, free will “in the common sense” do not exist: even though many futures are possible, a human being has no way to “control” which future will become true. The free will concept in the common sense is an illusion or subject feeling of the first observer that knows which future has just become true.
This conclusion of free will shakes the foundation of some of the moral systems. For example, it is hence unreasonable to blame a person who could have done better in his life for not doing his best, because he acts only as the first observer but not the chooser of his own life, and he did not actively chose to live a worse version. Punishment can hence only be justified as a filtering system for selecting more good entities from the past and also as a pressure to force more good in the future, but can no longer be justified in a way that assumes every one could have avoided being punished.
What is free will?
There are many different definitions of free will. I try to narrow the discussion using the following thought experiment:
If we assume that a common sense normal human-being has free will, then it at least means that this human being can choose to raise either his left arm or his right arm in the next minute. I assume that raising either arm will not have any positive or negative outcome for this human, i.e., the man is not threatened or tempted to raise either arm. I assume he has the ability to raise either or both arms, i.e. he is not handicapped in any sense. I also assume that his free will is “strong” enough to overpower his orientation habit, which means that it does not really matter if he is a left-handed or right-handed man. In short, if he raise an arm, it is only because he “chose” to do so, and this serves as an evidence that he has free will.
This setup has some important structure. Before the choice is made, there are (at least) two different possible futures: he raised his left arm or he raised his right arm. After the choice is made, the future pictures collapse into one possible future, and before the man exposes the choice (assuming there is no fancy brain activity monitoring cables on the man’s head), the man is the only person in the world that knows which one is likely to be the possible future. The view of the future is asymmetric for the chooser and a non-chooser of the matter. I call this asymmetric view of the future “the advantage of the chooser”.
As a simplified model, there is an exact time point in the procedure when the chooser made the choice. The question is how the choice actually comes out? Before that time point, the man did not choose, because the choice is not out yet. After the time point, the man also did not choose; the choice is already made. Right at that time point, the man made the choice, but how did that happen? At that time point, did the man have any sort of control on how the choice is made that is stronger than anyone else in the world?
My claim is that the choosing is completely an objective random outcome, probably derived from an observation of a quantum state in the man’s body (or some other inherently random entropy source). The chooser himself has no capability to control the result of the choice any more than any other men in the world.
How to prove that claim? Here is the proof:
Let’s build a robotic machine that has two arms and does one and only one thing: upon requests, it can raise one of its arms based on a random number generator that derives from some quantum state in the machine (or some other entropy source). From an outside observer’s point of view, this machine is no different from the choosing human that I mentioned above, in terms of having a free will or not. The choosing procedure is exactly the same, and there is no observable or measurable difference that shows the human chooser has more control on the choosing result than the machine chooser. Applying Occam’s razor, with no observable difference, the two entities should be considered equivalent. This means that if we admit that the machine does not have free will (because it seems too dumb to be able to choose), then we also have to admit that the human chooser also does not have free will; similarly, on the other hand, if we say that the human chooser has free will, then we also have to admit that the machine we just built also has free will.
Even when I am actually the human chooser myself, where I am totally aware of the choosing procedure and my mind activities, when I review the entire choosing procedure afterwards, I also have no clue where the choice comes from at the choosing point. I can observe no difference if the choice happen to be derived from the moving direction of an object, say a bacteria, in my brain.
In short, free will (if exists) is nothing more than a boxed machine equipped with a random number generator.
Why we strongly feel that we can choose?
So we cannot choose with free will, then why we strongly feel that we can choose?
First, most of the choices are largely algorithmic; they do not come from completely nowhere. Some come from reasoning, some come from external threatening or tempting, some come from our habits, some come from our feelings. These choices are largely the result of an execution of existing deterministic and predictable algorithms, though the algorithms might take random variables as inputs at some (critical) steps, which makes them not completely deterministic. The point here is that we do make choices by executing our own deterministic algorithm in our body, but since it is deterministic, it is not free will (in common sense), although it often gives us a strong feeling of being able to control the outcome during the execution. The input for the algorithms (including the algorithm itself, if you think of it a program) is often already determined. Just that this execution procedure is often the very first time that we really understand where this particular input leads us.
Second, for the non-deterministic part, we strongly feel that we can choose mainly because of “the advantage of the chooser”, because we are the first one that observes the result of the choice, and we can even “choose to” hide it if we want for many cases (again, whether we “choose” to hide it or not here either deterministically based on an the algorithm execution in our body, or non-deterministically based on another choice that derives from the entropy source in our body). It is this chooser’s advantage that distinguishes ourselves from all the other human-being existence in the world, and makes us individuals.
In short, we strongly feel that we can choose because we observe the non-deterministic result of “choice” as the first observer, and deeply embeds these choices in our (largely deterministic) execution of our own thoughts.
How it breaks our moral system?
Many concepts of our moral system is built on the assumption that human beings have free will. We think that a man is morally responsible for his past behaviors because he could choose on his own behavior. (I still remember the scene in Minority Report, where the oracle girl tells Tom Cruise that he can choose.)
However, if we admit that nobody can choose based on free will, a reasonable man can no longer blame a student for not working hard enough much more than blaming the result of a quantum state measurement, a reasonable man can no longer blame a criminal that kills innocents much more than blaming a boxed machine that equipped with a random number generator, a reasonable man can no longer blame North Korean or China not having enough human rights and democracy much more than blaming there is no life existence on the planet of Mars. The world is not deterministic, but it exists in its current form purely due to chance, where nobody has any control over it.
Bad things happen because it is unlucky.
How it won’t completely break our moral system?
Not everyone uses Occam’s razor for understanding the world, and the society is not organized purely based on reasoning, but also based on people’s belief. Just like people will continue their belief in God even there is no evidence for God’s existence; people will also continue their belief on free will (where a reasonable man that cannot blame them for believing their own believes, because they do not have any control over it.)
We Chinese often really care about how to succeed in career and life (and often success here often just means being both rich and happy, and sometimes only rich). For these matters, it is still true that if a person succeeds in its career, with high probability he worked hard and acted wisely. It is still true that if a person does not work hard and act wisely, it is unlikely that he will succeed. However, it is not true that a man succeeds because he chooses to work hard and act wisely; it just happened by chance. Also, it is not true that a man can always succeed if he will choose to work hard and act wisely; he has no control of choice over if he will work hard or not.
So if you are already working hard and acting wisely, you should keep doing it. Although you do not have free will to choose, it should not affect how you execute your deterministic algorithm to maximize the chances to have a happy life in the future if your algorithm is based on reasoning. All the reasoning and real-world causal relationship still remain unchanged. If you now think that it does not matter if you work hard or not, then you are understanding my view on free will incorrectly; it still matters. I just mean that you do not really have any sort of control over this thing.
If you are a man that want to live a life based on reasoning, a man that believe that an unexamined life is not worth living, then you should now understand that whether you can or will work hard or not, or succeed or not is purely by chance, where you have no control over it. The foundation of your moral system should be built based on accepting the reality and embracing all future possibilities and enjoying the entire life experience, rather than based on the existence of free will where you can choose your future.
In short, it only affects the moral system for those who want to live a life based on reasoning, which is not the case of all men.
Implication on building “free will” AI systems
Since free will (in common sense) is no different from a boxed machine with a random number generator inside, it is hence entirely possible to build a robot that has free will (or at least gives us feelings that it has free will). Free will is hence not a fundamental barrier for passing the Turing Test, although it still remains as hard as it is now to build a machine that has no observable difference from a human to an arbitrary observer (mainly due to the complexity of how human brains work).
Nobody make free will choices. The world exists as its current form purely by chance. Moral systems that are built on free will choices hence no longer hold their reasoning.
One of my friend feels sad after I convinced him that free will choices do not exist; I do not intend to break anyone’s world view and happy life, although I still happen to “choose” to write this down anyway. I sincerely hope that we can just carry on our own execution of our own algorithms for doing good (and not doing bad), and at the same time happily and peacefully accept and embrace all possible futures.