Is dualism true?

[harvey1]

Here's a paradox that seems that with today's brain scanning technologies one can envision how this paradox implies free will as well as dualism.

Imagine that you are the owner of a fantastic brain scanning machine that has recently been invented and is now harmlessly connected to your brain. The system is such that it can analyze the electro-chemical state of your brain, and based on that state can predict exactly what you will and must do next. Now, let's say that while sitting at the controls of this machine that it scans your brain upon pressing the green button and it comes back with, "you will press the purple button next." Now, upon hearing that you will press the purple button you decide to be a wise guy and you push the yellow button instead. The machine is wrong. But, how could it be wrong since it must know what your brain circuits would do upon hearing that you will press the purple button, and therefore the machine should be able to consider what your brain circuits would do even in that special case of knowing what you will do? If hearing that you would push the purple button, the machine must know that you would press the yellow button. However, if the machine told you that you would press the yellow button, then you would have surely not have pressed the yellow button. The machine must lie to you in order to predict your behavior. However, if it must lie to you, that means that it cannot predict your behavior by predicting your behavior. This suggests that there is no algorithm or scanning technology that the machine can use that predicts behavior when it has the task of reporting to you what your behavior will be. Therefore, the only way this could be true is if human behavior is indeterministic.

If human behavior is indeterministic, then wouldn't this mean that some form of dualism is true? That is, if no bridge laws exist that allow the machine to absolutely determine a human decision in all situations (as shown above), then the mental is not fully reducible to the physical. Dualism is the view that both the mental and physical exist, and existence is confirmed if the thing that is purported to exist cannot be explained in terms of other phenomena. Since the hypothetical machine cannot reduce every decision to a brain process that is scannable, wouldn't this suggest that there exists some non-physical component to the brain called the mind (i.e., dualism)?

[McCulloch]

Feedback.

The machine predicts what you would do, as if it had not given you its answer. If it tries to predict the influence of its own answer on your action, it could go into an infinite loop. It could, with all honesty, say that the button you press next is determined by which button the machine says you would press next and therefore, it cannot (or will not) tell you which button you will press.

I could build a device which presses one of two buttons using a pseudo-random number generator. A powerful enough computer with the same pseudo-random algorithm and seed could predict the device's actions. Now, in keeping with your thought experiment, I altered the device to take as input, the computer's prediction and alter the random number generator's seed based on that prediction. This has now rendered the prediction invalid, since it was based on the unaltered device's behaviour. However, the device's behaviour is still completely deterministic.

Let's take it a step further and upgrade the computer so that it can take the feedback into account. It's predictions, if the software is bug free and accurate, would either be accurate (it predicts a particular button and the device presses that button) or indeterminate (it computes that whatever it predicts, the device will behave in a way not to agree with the prediction). Have we given the device free will?

[harvey1]

It could, with all honesty, say that the button you press next is determined by which button the machine says you would press next and therefore, it cannot (or will not) tell you which button you will press. I could build a device which presses one of two buttons using a pseudo-random number generator. A powerful enough computer with the same pseudo-random algorithm and seed could predict the device's actions. Now, in keeping with your thought experiment, I altered the device to take as input, the computer's prediction and alter the random number generator's seed based on that prediction. This has now rendered the prediction invalid, since it was based on the unaltered device's behaviour. However, the device's behaviour is still completely deterministic.

Very true. However, as you somewhat alluded, this only pushes the issue up a notch. Instead of the Machine giving us a prediction of what button we will press, the Machine gives us a complete table of what button we will push if the Machine says press button X (e.g., purple button, red button, etc). So, for example, the table (provided to us by the Machine) shows that if the Machine says we will press "purple" then the table shows that we will actually press "yellow," if it says we will press "red," then the table shows that we will actually press "orange," and so on. In that particular case, once we have this table provided to us by the Machine, we can again foil the Machine even knowing what button we will press. For simplicity, let's call this a counterfactual decision table.

Now, let's say, as you suggested, that the Machine knows the "random" number generator inside our heads used to pick a button to press. (As you know there are no true random number generators--i.e., classical ones at least.) If we know the outputs of every calculation (which we can know for pseudo-random generators), then we can know everything the Machine knows. In that case, we can have access to this counterfactual decision table just like above, and then we can again foil every prediction by the Machine. This again suggests that we are not deterministic machines since we can always foil the Machine's attempt to know what button we will press.

Let's take it a step further and upgrade the computer so that it can take the feedback into account. It's predictions, if the software is bug free and accurate, would either be accurate (it predicts a particular button and the device presses that button) or indeterminate (it computes that whatever it predicts, the device will behave in a way not to agree with the prediction). Have we given the device free will?

It doesn't matter if the device is a computer or a human being. If the Machine cannot know the decisions of the device after providing this counterfactual decision table, then the device (or human) is indeterministic. In the case of a device however, there must be a circuit that uses a decision table or calculation (e.g., a pseudo-random calculation) to make a "decision," and therefore the Machine can always know what the computer would "decide"--even if it provided the contents of the counterfactual decision table. In the case of humans, this could not be the case since we can always decide differently than what a counterfactual decision table says will happen, and therefore we have free will. Since this couldn't be the case if the Machine could scan our heads for how we would always respond, I think this is a good reason to think that dualism is correct.

[McCulloch]

However, as you somewhat alluded, this only pushes the issue up a notch. Instead of the Machine giving us a prediction of what button we will press, the Machine gives us a complete table of what button we will push if the Machine says press button X (e.g., purple button, red button, etc). So, for example, the table (provided to us by the Machine) shows that if the Machine says we will press "purple" then the table shows that we will actually press "yellow," if it says we will press "red," then the table shows that we will actually press "orange," and so on. In that particular case, once we have this table provided to us by the Machine, we can again foil the Machine even knowing what button we will press. For simplicity, let's call this a counterfactual decision table.

Now, let's say, as you suggested, that the Machine knows the "random" number generator inside our heads used to pick a button to press. (As you know there are no true random number generators--i.e., classical ones at least.) If we know the outputs of every calculation (which we can know for pseudo-random generators), then we can know everything the Machine knows. In that case, we can have access to this counterfactual decision table just like above, and then we can again foil every prediction by the Machine. This again suggests that we are not deterministic machines since we can always foil the Machine's attempt to know what button we will press.

We can only foil the machine if the machine provides us with its predictions. That feedback loop invalidates the machine's predictions. If the machine printed its conterfactal decision table hidden from our observation, then if the machine was working properly, it would correctly predict our actions.

Let's take it a step further and upgrade the computer so that it can take the feedback into account. It's predictions, if the software is bug free and accurate, would either be accurate (it predicts a particular button and the device presses that button) or indeterminate (it computes that whatever it predicts, the device will behave in a way not to agree with the prediction). Have we given the device free will?

It doesn't matter if the device is a computer or a human being. If the Machine cannot know the decisions of the device after providing this counterfactual decision table, then the device (or human) is indeterministic. In the case of a device however, there must be a circuit that uses a decision table or calculation (e.g., a pseudo-random calculation) to make a "decision," and therefore the Machine can always know what the computer would "decide"--even if it provided the contents of the counterfactual decision table.

If the device were constructed to always press a different button than the one predicted, then the Machine could never correctly predict the device's action. That does not prove dualism for the device.
It is only a variation of the paradox of the card with the words "The statement on the other side of this card is false." on one side and "The statement on the other side of this card is true." on the other. You cannot determine the validity of those statements. It seems to me that you have proven that my card has free will.

In the case of humans, this could not be the case since we can always decide differently than what a counterfactual decision table says will happen, and therefore we have free will.

Only if we know and can be influenced by the information on the conterfactual decision table.

[harvey1]

If the device were constructed to always press a different button than the one predicted, then the Machine could never correctly predict the device's action. That does not prove dualism for the device. It is only a variation of the paradox of the card with the words "The statement on the other side of this card is false." on one side and "The statement on the other side of this card is true." on the other. You cannot determine the validity of those statements. It seems to me that you have proven that my card has free will.

Not quite. This is true in the original OP where I was laying out the paradox. However, the paradox doesn't vanish as is clear below when we move up a notch. Here's how it works:

1) Scanning Machine knows our actions unless it tells us in advance what we will do
2) If the Machine tells us in advance what we will do, then we will do the opposite
3) "If the [human] always press[ed] a different button than the one predicted, then the Machine could never correctly predict [our] action."
4) In any case, the Machine would possess a counterfactual decision table that would know (3) as long as it did not provide the counterfactual decision table to us
5) If the Machine provided the counterfactual decision table in advance, then we will not follow the table and the Machine would not possess a table to predict our response
6) (5) is not true for a physical device since the counterfactual decision table would reflect the complete circuit (or algorithm) of the device, therefore the device would have to follow the counterfactual decision table of the Machine
7) Hence, humans can decide indeterministically, while machines are deterministic and predictable

Note that when we move up a notch in (4) and (5), there is a difference between your card example (or computers, devices) and humans. In the case of the card example, there is a simple counterfactual decision table that exists that the card cannot violate.

In the case of humans, this could not be the case since we can always decide differently than what a counterfactual decision table says will happen, and therefore we have free will.

Only if we know and can be influenced by the information on the conterfactual decision table.

That's a big IF. That means there exists an exception to determinism. When we accept that there's no counterfactual decision table known by the Machine that it can use to predict our actions (even though the Machine for all practical purposes is omniscient), then our decisions must be capable of being indeterministic thought, and irreducible to the material structures of the brain. The only way this wouldn't be the case is if we just couldn't violate the counterfactual decision table, in which case we would find ourselves strangely unable to do differently than what we were told what we would do. Since this is not a tenable situation, we can rule out determinism as the basis of our decision making.

[McCulloch]

1. Scanning Machine knows our actions unless it tells us in advance what we will do
2. If the Machine tells us in advance what we will do, then we will do the opposite
3. "If the [human] always press[ed] a different button than the one predicted, then the Machine could never correctly predict [our] action."
4. In any case, the Machine would possess a counterfactual decision table that would know (3) as long as it did not provide the counterfactual decision table to us
5. If the Machine provided the counterfactual decision table in advance, then we will not follow the table and the Machine would not possess a table to predict our response
6. (5) is not true for a physical device since the counterfactual decision table would reflect the complete circuit (or algorithm) of the device, therefore the device would have to follow the counterfactual decision table of the Machine
7. Hence, humans can decide indeterministically, while machines are deterministic and predictable.

Note that when we move up a notch in (4) and (5), there is a difference between your card example (or computers, devices) and humans. In the case of the card example, there is a simple counterfactual decision table that exists that the card cannot violate.

(6) is not necessarily true. Let's say that there are six buttons. My device has a pseudo-random number generator and it is programmed to randomly select one of the five buttons not predicted by the Machine. The Machine has complete and perfect knowledge of the device, yet the Machine cannot accurately predict which button the device will select if the prediction is provided into the device. My admittedly deterministic and predictable device thwarts the predictive abilities of the all knowing Machine. Yet with the exact same set-up, you assert that a human in the same place as my device has free will as demonstrated by this experiment.

In the case of humans, this could not be the case since we can always decide differently than what a counterfactual decision table says will happen, and therefore we have free will. [...] The only way this wouldn't be the case is if we just couldn't violate the counterfactual decision table, in which case we would find ourselves strangely unable to do differently than what we were told what we would do. Since this is not a tenable situation, we can rule out determinism as the basis of our decision making.

Your knowledge of the existence or the content of a counterfactual decision table, I have shown from the device example, invalidates the predictions in the table, without removing determinism from the device.

Is it possible to deliberately do that which you have not decided to do? You seem to be arguing the affirmative.

[harvey1]

(6) is not necessarily true. Let's say that there are six buttons. My device has a pseudo-random number generator and it is programmed to randomly select one of the five buttons not predicted by the Machine. The Machine has complete and perfect knowledge of the device, yet the Machine cannot accurately predict which button the device will select if the prediction is provided into the device.

But, why wouldn't the Machine know exactly which button the device will select if the Machine has the algorithm and can run the same calculation as the device? For example, let's say that the device uses as its argument [6] which is the Machine's answer to what the device would choose if the Machine said nothing. The device then "randomly" selects buttons [1]-[5]. The counterfactual decision table is based on this same pseudo-random calculation already performed by the Machine (i.e., based on the state of the device at t). The Machine knows the result of that pseudo-random calculation because it has the entire state of the device. So, for example, if the device does its calculation at t based on [6] and the calculation comes back as [1], then the counterfactual decision table will show [6, 1]. Notice how this is recursive if the device has access to the counterfactual result. If the device uses [6,1] to do its pseudo-random calculation and arrives at [5], then the Machine's counterfactual table for the device will show [6,1,5]. This arms race stops conceptually at infinity, in which case neither the Machine nor the device know which button will be pressed by the device. (Such a device would be indeterministic since the Machine cannot predict the botton it will press--the device's choice is uncomputable.) Any finite calculating device, however, would only make so many recursive step calculations, and the Machine would know how many steps that is (based on the contents of its circuits), and therefore the Machine will always be able to predict the devices choices.

In the case of humans this is not the case. The human only needs to know what the prediction of the Machine is (no matter how many recursive steps it computes), and then violate that choice. Notice that this option is not available to the finite device since the Machine calculates what the device will choose (even if the device is trying to violate the Machine's last prediction). If the Machine doesn't know, then it's not a finite calculating device, in which case off to infinity they go. Which, again, means that the device's decision is uncomputable.

My admittedly deterministic and predictable device thwarts the predictive abilities of the all knowing Machine. Yet with the exact same set-up, you assert that a human in the same place as my device has free will as demonstrated by this experiment.

Only a device performing an infinite number of computations can outwit the Machine. If such a device existed, then it would be indeterministic by definition since its choice cannot be determined even in principle. That would be some device.

Is it possible to deliberately do that which you have not decided to do? You seem to be arguing the affirmative.

My argument is only to show that in any scenario where the Machine scans our brain and tells us what we must decide that we can violate the prediction of the Machine. Hence, brain structure alone is not enough to determine what we will do (or dualism has some important truth to it).

[McCulloch]

(6) is not necessarily true. Let's say that there are six buttons. My device has a pseudo-random number generator and it is programmed to randomly select one of the five buttons not predicted by the Machine. The Machine has complete and perfect knowledge of the device, yet the Machine cannot accurately predict which button the device will select if the prediction is provided into the device.

But, why wouldn't the Machine know exactly which button the device will select if the Machine has the algorithm and can run the same calculation as the device?

Because the Device has been programmed to never select the button predicted by the Machine. If the Machine predicts [6] then it knows that the Device will not select [6]. It then recalculates and determines that if the Machine predicted [6] then the device would select [1]. There is no need for infinite recursion.
At a particular state t if the Machine predicts:

1. the Device will select [3]
2. the Device will select [5]
3. the Device will select [1]
4. the Device will select [2]
5. the Device will select [6]
6. the Device will select [5]

The Machine cannot accurately predict the Device even though the Machine has perfect knowledge of the Device's processes and algorithms. The Device's behaviour is completely deterministic.

I have shown that a device with less than an infinite number of computations has outwitted the Machine.

Now put a human in the place of the Device. If the Machine does not convey the prediction to the human, the Machine accurately predicts the human's behaviour. But if the Machine lets the human know its prediction, that prediction will affect the human's behaviour. If at time t² the Machine predicts:

1. the human will select [3]
2. the human will select [4]
3. the human will select [5]
4. the human will select [1]
5. the human will select [3]
6. the human will go for a coffee.

The Machine cannot accurately predict the human even though the Machine has perfect knowledge of the human's processes and algorithms. The human's behaviour is completely deterministic.

Is it possible to deliberately do that which you have not decided to do? You seem to be arguing the affirmative.

My argument is only to show that in any scenario where the Machine scans our brain and tells us what we must decide that we can violate the prediction of the Machine.

Yes, but so can a relatively simple deterministic device.

Hence, brain structure alone is not enough to determine what we will do (or dualism has some important truth to it).

No, not yet. At least not by this thought experiment.

[Bugmaster]

As you know there are no true random number generators--i.e., classical ones at least.

I'm still reading through the thread, but I just wanted to jot down a quick note on this.

True random number generators do exist. Radioactive decay, electronic noise, and weather patterns are all random processes; they can be harnessed to provide a stream of truly random numbers. Actually, a much simpler device, consisting of ping-pong balls with numbers on them, has been in use for many years now in order to generate winning lottery tickets. It would be fairly simple to use these random numbers as an input to a computer program; in fact, people have done so already.

I think what you meant to say was, "there are no perfect pseudorandom number generators".

Also, keep in mind that determinism could be false, but dualism could still be false as well. Anyway, I'll keep reading the thread.

[Bugmaster]

Ok, I've finished reading the thread.

I think McCulloch is doing an excellent job of defending his argument in a deterministic context. So, I'll take a different approach.

Let's say we hook up a radioactive sample (of radium, or something) to a geiger counter, and hook the geiger counter up to a microchip, which in turn will drive the button-pressing mechanism. The chip will count the number of clicks the geiger counter emits, per minute (N); it will then press the button (N % 6) + 1, every minute.

Since radioactive decay is truly random, no Machine in the universe would be able to predict which button our device will press. Does this mean that our device has free will ?