Tuesday, May 26, 2020

Perdurance, physicalism and relativity

Here is a very plausible thesis:

  1. Exactly one object is a primary bearer of my present mental states.

This is a problem for the conjunction of standard perdurance, physicalism and special relativity. For according to standard perdurance on physicalism:

  1. The primary bearers of my mental states are time slices.

Now consider all the time-slices of me that include my present mental states. There will be many of them, since there will be one corresponding to each reference frame. On relativistic grounds none of them is special. Thus:

  1. Either all or none of them are the primary bearers of my present mental states.

If all of them are the primary bearers of my present mental states, we violate 1. If none of them are, then there is no primary bearer of my present mental states by 2, which also violates 1.

28 comments:

Mike Rathbone said...

Speaking of mental states. If all points on a timeline are equally real, how come I only know t1 ("now"), t0 (the past), but not t2 ("the future")? If I am equally present at all points, how come I only have knowledge of two of those points?

Alexander R Pruss said...

You know t1 and t2. But you know t1 at t1 and t2 at t2.

Walter Van den Acker said...

Alex

Yes, but why do you know t1 at t2 and not t2 at t1?

Alexander R Pruss said...

Walter:

Because your states at later times tend to be caused by your states at earlier times.

Though, actually, we know as much about the future as about the past. https://alexanderpruss.blogspot.com/2008/05/knowledge-of-future.html

IanS said...

Alex:

Isn’t the ‘special’ reference frame for my mental states the one in which my brain is at rest? Similarly, if my friend were in a spaceship travelling (relative to me) at close to the speed of light, the natural reference frame for his mental states would be one in which his brain was at rest.

wrf3 said...

This argument fails for computers, so why should we think it applies to humans?

I'll grant you #1, but #2 fails immediately. The state of a computer is not primarily time slices. If you use the Turing model of computation, the state is encoded by the state register, the instruction at that state, and the contents of the memory.

Since #2 fails, #3 fails.

Alexander R Pruss said...

Ian:

Different bits of my brain move at different speeds. Which one defines the reference frame?

wrf3:

The state register, the instruction at that state, and the contents of the memory are all parts of the time slice.

That said, while perdurantists say things like #2, I agree with you it's false. In the computer case, the problem is that the state is encoded by the state register, the instruction, and the contents of memory *when these are all validly synchronized by the system clock*. And being synchronized by the system clock is not something that happens on a time-slice, first because the clock is defined by a transition and, second, because the clock pulse propagates at a speed slower than light.

wrf3 said...

*when these are all validly synchronized by the system clock*

Well, yes, for architectures that use clocks for synchronization. But that's just a design choice, not a fundamental feature of computing. Our brains don't have a system clock, nor do dataflow architectures. Since a system clock isn't fundamental to computing, (Turing machines don't have a clock) it really shouldn't be a consideration in this argument.

Alexander R Pruss said...

I think that even in a dataflow architecture, you still have issues of validity of data and these issues involve time-dependence. Whether an electrical potential in the output of a logic gate means anything--and hence whether it is a part of the computational state--depends on whether we've waited out the gate delay from the time that the inputs arrived at the gate. So it's still true that the time slice is insufficient to define the computational state.

Pretty much any digital system will suffer from a version of this. There are probably parallels to the problem in analog systems, but I know very little about analog computation.

A Turing machine abstracts from all this by assuming that its states occur at specified discrete times at which all the states are defined. But the specification of these times will suffer from all these issues (namely, the times need to be so chosen as to ensure that there has been enough time for all the signals to have propagated, all the flip-flops to have settled, all the gates having finished their operations, etc.; and whether this specification has been met depends on the past).

wrf3 said...

Let me switch examples away from computers to something a bit more familiar. It doesn't change the argument, but may make it easier to follow. Consider a symphony. Yes, there's a beat, but the beat can change. The conductor might even stop his beat to allow for improvization. One section of the orchestra might be following one beat, another section another beat. An musician might be resting on a beat; another musician might be starting a note mid-beat; another musician might be holding a note across multiple beats.

So the primary states are not the time slices. And it should be obvious as to why. The "primary state" is the ensemble at a given (continuous) time. Now, some ensembles are simple enough that we can capture their state (think a CD of a recording of a performance). It is then no problem to play that CD in London as well as on the way to Alpha Centauri. Of course, this only captures the sound waves, it doesn't capture the motion of the musicians. But we could make a video. But this wouldn't capture the air moving through the instruments or the bows across the strings. Some things are too complex to duplicate exactly. Therefore, the only reason we think there is a "primary" us is because we don't yet have to technology to digitize the brain's ensemble (which, really, includes our entire body). And, even if we could, humans are neural networks that respond to external factors. The moment the external factors differ between clones, the clones will start to differ. That's why differences among Cylon models existed.

This whole argument, is, if you'll forgive me for saying so, "not even wrong."

Alexander R Pruss said...

BUt the ensemble at a single time doesn't have any sounds in it. At any given time, all there is the air pressure at various points in space. Compare the CD: at any given time, all we have are two 16-bit numbers (corresponding to the two channels). The two 16-bit numbers do not encode a sound.

wrf3 said...

BUt the ensemble at a single time doesn't have any sounds in it. At any given time, all there is the air pressure at various points in space.

Which is part of the ensemble. And sound is the air pressure at various points in space. Or is this a possible point of disagreement?

The two 16-bit numbers do not encode a sound.

Sure. Those 16-bit numbers could be, say, a color value, or one Unicode character, or... So what? In this case, they are part of an ensemble where one or more pieces of the ensemble are missing. But that's an accident of not being able to fully clone the actual ensemble. I don't see how it helps your argument.

I suspect we have a fundamental disagreement over the nature of thought. I will argue that thought is matter in motion in certain patters; you will argue that this cannot be so. That's a bigger argument that I'm happy to debate. Aristotelian arguments, such as those advanced by Feser and Searle, are demonstrably wrong.

Still, is part of our disagreement over whether or not a falling tree in a forest makes a sound absent an observer? I say "yes", you say "no"?

Alexander R Pruss said...

I think sound is the change in air pressure at one location. But change is not something you can discern from a snapshot of the universe at one time. Similarly, if thought is matter in motion, then one cannot read off a thought from the state of the universe at one time, because one cannot read motion off from the state of the universe at one time.

wrf3 said...

But change is not something you can discern from a snapshot of the universe at one time...

Really?

Michael said...

wrf3,

The point I'm assuming you're trying to make is that from a snapshot of the universe we can tell effects of change, and infer change.

If this is true, could you explain from that how you know, from the picture itself, how that picture is of a real bullet through an apple and not a picture of a very well made suspended sculpture (that would be still during the picture-taking)?

To me it seems to require background information, and that background information would be from another time.

Alexander R Pruss said...

wrf3:

Well, imagine an instantaneous photo of a billiards table in the middle of a game. Can you tell whether the balls are moving or not?

wrf3 said...

Alex: Well, imagine an instantaneous photo of a billiards table...

No, but that's because a photo is an incomplete snapshot of the state of the system. If the mass and the velocity of each ball at the time the picture was taken were included in the snapshot, I could.

wrf3 said...

Michael: If this is true, could you explain from that how you know, from the picture itself, how that picture is of a real bullet through an apple...

I will argue that the snapshot of state is more than just the picture, it's also the description of the setup.

To me it seems to require background information, and that background information would be from another time.

Sure. It requires information up to the time the picture was taken. But our brains record historical data. So the system in this case is the observer, the apple, the gun, the bullet, and the camera. The camera records one aspect of the system at t0; the observer at t0 another (snapshots can, after all, be the compilation of information from different observers. Cameras, thermometers, audio recorders, people, ...)

Michael said...

wrf3,

I misunderstood you (and Alex,) as referring to just a snapshot of position and possibly other non-change parameters. If we have for example, a snapshot of velocity or another instantaneous measure of change, then yes I think we can read that off, along with our history of understanding on it.

Alexander R Pruss said...

wrf3:

But velocity is rate of change of position. And so velocity is defined in terms of what happens at other times.

wrf3 said...

Alex: velocity is defined in terms of what happens at other times.

So what? I really hope you take the time to answer this "so what?" because I'm truly baffled by your responses. There's a point of disconnect between us, and I really want to (try to) understand what it is and why.

First, systems with memory are different than systems without memory. The state of a system at Tn can include the descriptions of the states of the system at Tn-1, Tn-2, ... (which is one reason why there is one, and only one, instance of your state at Tn, since the state at Tn+1 will include information about Tn).

Second, the state of a system at Tn is the system. Descriptions D of the system at S at Tn are necessarily incomplete, because D is a different physical system from S. That is, there are very few instances where the description of a thing is that thing.

Alexander R Pruss said...

The target of my post is perdurantism. Perdurantists seem to believe that fundamental changeable properties depend only on what happens at a single time ("on a slice"). So views of systems on which the state of the system at a time t includes information about what the system was like at other times are not going to be relevant here.

IanS said...

The state of a system, as usually understood in non-quantum physics, includes the momenta. This has always worried me a bit, for the reason you give. But worries aside, it seems necessary to make the physics work. If physicists’ states can include momenta, maybe perdurantists’ time-slices can too.

The temperature of an object is related to the motions of its atoms. Must a perdurantist say that it has no temperature? Or else that temperature is not a ‘fundamental’ property?

Note that in QM the state is described by a wavefunction, which implies distributions of both position and momentum. This seem to dissolve the problem. Of course, it brings different problems.

wrf3 said...

Alex: The target of my post is perdurantism...

Thanks for taking the time to further answer my questions.

However, I just can't get past the idea that pedurantism is just ... silly. It ignores what we know about physics. It makes no sense to say So views of systems on which the state of the system at a time t includes information about what the system was like at other times are not going to be relevant here. Entangled quantum states include information about past history. You can't get around it.

It also doesn't make sense to talk about copies of "mental states". The moment you measure something in order to determine what the state is, you change the state. That's why making exact copies of objects is impossible (see the quantum No-Cloning theorem).

This further underlies the problem of how they test for equality. What, exactly, is meant by "equal mental states"? Is it the state of the brain considered as a classical system? A quantum system? A non-physical system? If it's non-physical, how do you measure it to test for equality?

Alexander R Pruss said...

Ian:

The tentative conclusion I've come to about momenta is that in classical physics they are fundamental physical quantities and that the thesis that are equal to the the mass multiplied by the time derivative of the position is actually a law of nature rather than a definition of the momentum. Here's one reason to think this. It should be possible for a Newtonian universe to begin with an initial moment of time. But if it does that, then if momenta are defined the usual way as derivatives, they depend on what happens *after* the initial moment of time. And that would make the evolution of the system involve something too similar to backwards causation.

But now if momenta are fundamental physical quantities and it is a law of nature that they have the standard connection to time derivatives of position, then even if the state of the system at time t does include the momentum, the state of the system does not include what happens at other times. For it is metaphysically a contingent fact that the system follows these laws of nature, so there will be a metaphysically possible world which includes the same state at t but where the particle having the non-zero momentum at t doesn't actually move anywhere during an interval of times centered on t.

That's classical physics. But in quantum physics, the present state does not encode sufficient information about the past, since the present state could have just resulted from an extremely unlikely indeterministic transition from some weird state that has no people in it, say.

So, in the comments above I was making a simplified version of this argument: Either momenta are defined in terms of time derivatives of position or not. If so, then slices don't have momenta, etc., and hence they are insufficient for mental properties. If not, then slices do have momenta, but their having these momenta is insufficient to metaphysically guarantee anything about the past or future, and hence slices are still insufficient for mental properties.

Alexander R Pruss said...

wrf3: Quantum states only carry probabilistic information about what happens at other times. They don't carry information about what *definitely* happened at other times.

What is not possible is a physical copying process. It seems quite obvious that God, if he exists, could duplicate a state. Since the duplicating isn't a physical process, it wouldn't violate the theorem.

wrf3 said...

Alex: And that would make the evolution of the system involve something too similar to backwards causation.
Wait, what? At the initial moment of time the universe starts expanding. Conservation of momentum is a result of the homogeneity of space (cf. Noether's theorem). So, like a puck sliding on ice, it just depends on the initial impulse (ignoring friction and gravity).

the state of the system does not include what happens at other times...
Define which system. The nature of space itself defines what's going to happen (in the classical sense).

If so, then slices don't have momenta, etc....
And Zeno's arrow will never reach the target. Since it does, either there's something wrong with your argument or we're all hallucinating.

wrf3 said...

Alex: They don't carry information about what *definitely* happened at other times.
Oh, but they do. Entangled particles carry information that the two particles were definitely entangled in the past. That means if you measure one entangled particle as, say, spin up then you will measure the other particle as spin down.

If God were to make a non-physical duplicate of me, how would you distinguish between the me's? How would you distinguish me and God? After all, God has no parts. This is a far more interesting problem than perdurance... because if you say that God is the primary bearer of my mental states (cf. Acts 17:28) then there are no time slices (because God does not change, Jam 1:17), and so the question doesn't even get off the ground.