Relativity Explained Via Movement, Not Time

[Dale]

Events.

Well, it can't just be any events since some events are space-like separated, and even for timelike separated events I can think of examples where you wouldn't call the resulting count time.

 

[Selraybob]

It is events. When I say "Time is a count", it's really saying that since the first caveman looked at the moon and grunted out something like, "moon above again," and then made a mark on the wall, and then another the next time the moon was above, is counting up some events and calling that Time. It's the counting that comes first and the definition of Time that comes after. You can count moons moving around the Earth or suns appearing to move around the Earth or the ticks of a grandfather clock or water clock or atomic clock. But for comparing our lives against, and meeting someone down at the dock, it's good to have some sort of counter that's pretty regular. Doesn't matter what it is though.

 

[Dale]

It is events. ... Doesn't matter what it is though.

So counting ticks on my Geiger counter is time. If I put my Geiger counter in a lead box time stops. Is that really what you meant, because that is what you said.

Here is another set of events, the tick marks on my ruler when illuminated by a specific of light. The count is 12. So time is 12. Doesn't sound like your definition is related to the usual idea of time.

 

[my_wan]

Well, it can't just be any events since some events are space-like separated, and even for timelike separated events I can think of examples where you wouldn't call the resulting count time.

True, which is exactly what gets us into trouble with simultaneity, which I even recently embarrassed myself with. Yet even then the one constraint that allows us to define the relativity of simultaneity is the constraint that a 'local' event set sequence, E_i --> E_j --> E_k, cannot place any event in the set both before and after some other event in that set. That would amount to time travel.

Even if we assumed some 'toy' model of purely mechanistic events such as molecular collisions, if the background is undefined then local event sets must disagree on clock rates in many different circumstances.

 

[Dale]

True, which is exactly what gets us into trouble with simultaneity, which I even recently embarrassed myself with. Yet even then the one constraint that allows us to define the relativity of simultaneity is the constraint that a 'local' event set sequence, E_i --> E_j --> E_k, cannot place any event in the set both before and after some other event in that set. That would amount to time travel.

Even if we assumed some 'toy' model of purely mechanistic events such as molecular collisions, if the background is undefined then local event sets must disagree on clock rates in many different circumstances.

So then what sets of events can you count and call the resulting count "time"?

 

[my_wan]

So counting ticks on my Geiger counter is time. If I put my Geiger counter in a lead box time stops. Is that really what you meant, because that is what you said.

Here is another set of events, the tick marks on my ruler when illuminated by a specific of light. The count is 12. So time is 12. Doesn't sound like your definition is related to the usual idea of time.

If you insure that those ticks are scaled properly yes, which is why we can do carbon dating. Yet like the linear discontinuities you can get from any given pair of individual clicks we have the uncertainty principle telling us we are getting these small scale discontinuities at a fundamental level also. But averaged over they go away.

 

[Dale]

If you insure that those ticks are scaled properly yes

OK, so now we are no longer counting, but counting and scaling our count. That is fine, how do we determine the correct scaling factor for the count?

 

[my_wan]

OK, so now we are no longer counting, but counting and scaling our count. That is fine, how do we determine the correct scaling factor for the count?

Relative to any other local event set with has a reasonably stable linearity. The most fundamental of which is apparently a Planck unit of time, which is subject to the Uncertainty Principle. If you arbitrarily presume some mechanistic basis then what makes it linear does not mean linear in any absolute sense, it just means linear with respect to the mean local fundamental event sets defining the system. After all, relative to a clock under an acceleration of proper motion, linearity in terms of time alone loses relevance.

 

[my_wan]

The scaling is required simply so that our choice of event sets to count time is independent of the particular set of events used to do the counting.

 

[Dale]

Relative to any other local event set with has a reasonably stable linearity.

So the tick marks on a ruler is a good example since it is stable and very linear.

 

[Dale]

The scaling is required simply so that our choice of event sets to count time is independent of the particular set of events used to do the counting.

That doesn't make any sense at all if time IS the count. How can time be the count of events and yet be independent of the events counted?

 

[my_wan]

That doesn't make any sense at all if time IS the count. How can time be the count of events and yet be independent of the events counted?

It is not independent of the count, which is the point. But if you have an event set, and many different ways of combining those set and subsets, then to get any meaningful count they all must be given a consistent scale, even though nature is scale independent in general. The reason they can be scaled together would be because they are are mere regrouping of the same sets.

 

[Dale]

It is not independent of the count

You just said that you needed to scale it so that it was independent.

to get any meaningful count they all must be given a consistent scale

And exactly what is it that must be consistent about the scale? In other words, you have two valid sets of events to count and you scale one by a factor of two in order to achieve consistency, what is it about the events that required a factor of two scaling to achieve consistency, and how can you know that you need a factor of two and not three?

 

[my_wan]

You just said that you needed to scale it so that it was independent.

I said independent of our "choice" of event sets, not independent of the events themselves.

And exactly what is it that must be consistent about the scale? In other words, you have two valid sets of events to count and you scale one by a factor of two in order to achieve consistency, what is it about the events that required a factor of two scaling to achieve consistency, and how can you know that you need a factor of two and not three?

Two parts:
1: "events that required a factor of two"
If you have a single finite superset of events and you have a measuring device that measures ticks for ever pair of events, then that MUST relativistically scale as 2 times every event. How much linearity there is between the individual events is immaterial, like claiming you can stop time for some specified period of time. Even if it we overlook the silliness and presumed it did occur it has NO empirical meaning.

2: "how can you know that you need a factor of two and not three"
Because the clock that counted one event for every two events would not be linear by any measure other than a 2 to 1 ratio. This ratio is what defines what we can measure, not the naked quantity in itself. That would require absolutes like in Newtonian physics.

 

[Dale]

Unfortunately, I will not be able to continue the conversation tonight. I thought it would take fewer iterations than it did.

The point is that this definition of "time is counting events" is not sufficient. There are some events which counting them doesn't measure time because the events are separated by space, not time. There are other events which counting them does not measure time because they are at random intervals. For still other sets of events you have a good feeling that they are measuring time, but you have to scale your counts in order to say that they are all measuring time. So for all of these reasons there is more to measuring time than simply counting events.

All of these objections can be addressed, but to do so requires that you identify something about the various counts beyond the mere counting itself. The very fact that you can use counts of a variety of sets of events (properly scaled) to measure time indicates that there is something else besides simply the counting.

Unless you already have some concept of time already then there is no way to distinguish between spacelike sets of events and timelike sets of events. Unless you already have some concept of time then there is no way to determine if the scaling is correct.

What the counts and their scaling do is to define a unit of time, not to define time.

 

[danR]

I don't know how this post 'relativity explained by movement, not time' got so far off track.
You got movement, you got:

(amount of) space/time.

Time will always be around in the physics-conversation.

 

[my_wan]

Unfortunately, I will not be able to continue the conversation tonight. I thought it would take fewer iterations than it did.

I am accustomed to certain people, including physicist, who seem to grasp it straight away as though it was too obvious to mention, and others who find it very difficult.

The point is that this definition of "time is counting events" is not sufficient.

Is actually is sufficient. In fact from a purely empirical perspective some form of event counting is ALL we have. The derivation of the relativity of simultaneity is possible exactly because we can count event rates of external observers as measured by our own evet rate counters and compare the compare the incongruities. If time existed independently of these event counts why would the difference in events counts result in an actual difference in time rates per Lorentz? The one issue that imposes any other constraints that might qualify as indicating "count" is not "sufficient" is the extra constraint imposed by causality itself, i.e., the same event cannot occur both before and after a reference event at a point in space.

There are some events which counting them doesn't measure time because the events are separated by space, not time.

In those situations where the event count does not match your own local event count then you are measuring is not your own event count but somebody else's. Which we know to be real differences in real time else you could not come back younger than your own kids. The separation by space rather than time is moot by the inverse relation between space and time such that empirically they are not separate quantities.

There are other events which counting them does not measure time because they are at random intervals.

The randomness of an interval does not invalidate the linearity of those intervals on average. The rate of molecular collisions also have highly random intervals. Yet in order for Gibbs ensembles to have any empirical meaning they must average to a constant if equilibrium is maintained. This is the physical basis of emergent gravity theories based on relations defined by Brustein and Hadad and others where equilibrium is not maintained.
http://arxiv.org/abs/0903.0823

Randomness of intervals do not have any relevance to the constancy of average intervals one way or the other, but it would have relevance to the certainty with which we could, even in principle, determine a precise location in space or time relative to any observer.

For still other sets of events you have a good feeling that they are measuring time, but you have to scale your counts in order to say that they are all measuring time.

If I locally measure an inertial string and say it has a lengths of X and you measure it locally and say it has has a length of Y, why must we scale X and Y so to be equivalent? Because we are measuring the "same" string. Same thing if we are both measuring the "same" general set of local events. How do you treat background independence by any other means?

So for all of these reasons there is more to measuring time than simply counting events.

For all those same reasons, plus background independence causality, "counting events" is completely "sufficient".

All of these objections can be addressed, but to do so requires that you identify something about the various counts beyond the mere counting itself.

Yes, they can be enumerated as such:
1: Causality - The same event A cannot occur 'both' before and after event B at any given point in space-time.
2: Background Independence - The rate of a non-observable is empirically moot irrespective of what role it plays in the theoretical framework.
3: Relational, i.e., Relativity - The only value we can measure are not naked values, rather ratios which we assign a specific value to as though our local base event sets defined absolutes.

The very fact that you can use counts of a variety of sets of events (properly scaled) to measure time indicates that there is something else besides simply the counting.

Yes causality constraints. Yet the fact that locally the event sets being measured in for all practical purposes the same set of events requires that both measures scale together. Hence the scaling requirement is a physical requirement of consistency. Not some magical entity that exist in a matter free background.

Unless you already have some concept of time already then there is no way to distinguish between spacelike sets of events and timelike sets of events.

Absolutely true and well established physics, which is precisely why we get the so called clock paradox in SR. We can only make such distinction locally, due to the fact that we are essentially measuring the same space-like and time-like sets. Once you consider other points in space-time what you say here about the event count ideology is exactly true and well defined by Relativity. You can only pretend that space and time-like intervals are unique in pair of reference frames by assuming that one of those frames has a special status in its capacity to define those distinctions.

Unless you already have some concept of time then there is no way to determine if the scaling is correct.

If a ruler and a string were all that existed in the Universe how do you know the scaling of the ruler is correct. With background independence (also called coordinate independence) 'correctness' of a scale makes no difference whatsoever. We know that the ruler can linearly scale to itself and we know the ratio between the ruler and the string, what else empirically matters? How big the ruler and the string both are is a physically absurd empirically moot question.

What the counts and their scaling do is to define a unit of time, not to define time.

So if a Hilbert space is a mathematical tool with no 'real' physical significance, what makes the mathematics of defining time so special that somehow it is not only more real but 'real' independently of the intervals we measure? (That is not saying it is dependent on our choice of interval sets.)

 

[atyy]

@DaleSpam: @my wan: Is the difference between your points of view that physics is always the study of change dy/dx, but if there is only one "dimension" DaleSpam would call that "space" and my wan would call it "time"?

 

[Dale]

In those situations where the event count does not match your own local event count then you are measuring is not your own event count but somebody else's.

No, the count of a set of spacelike separated events is not measuring anyone's time.

For all those same reasons, plus background independence causality, "counting events" is completely "sufficient".

Yes, they can be enumerated as such:
1: Causality - The same event A cannot occur 'both' before and after event B at any given point in space-time.
2: Background Independence - The rate of a non-observable is empirically moot irrespective of what role it plays in the theoretical framework.
3: Relational, i.e., Relativity - The only value we can measure are not naked values, rather ratios which we assign a specific value to as though our local base event sets defined absolutes.

Yes causality constraints.

You are contradicting yourself saying that "counting events" is completely sufficient and then adding constraints. The "time is counting events" definition is clearly insufficient as you have de facto conceded. You must add such enumerated constraints or clarifications as above.

The causality constraints are particularly problematic for a definition of time. Causality requires a notion of time, so putting a causality constraint into your definition of time makes the definition circular. In particular, without a definition of time how do you determine "before" and "after"?

Btw, I am not opposed to a background independent relational definition of time. I just think that "time is counting events" is way oversimplified. Something this important needs quite a bit more effort and care than that. Many things which qualify as counting events do not qualify as measuring time.

 

[Dale]

@DaleSpam: @my wan: Is the difference between your points of view that physics is always the study of change dy/dx, but if there is only one "dimension" DaleSpam would call that "space" and my wan would call it "time"?

I have not asserted a point of view here, I am only pointing out that "time is counting events" does not work. I believe that I understand my_wan's motivation, I think he wants to avoid the Newtonian idea of time as some undetectable thing which flows and instead focus only on the physical observables. I agree with that goal; I just think that his approach is not right.

 

[atyy]

I have not asserted a point of view here, I am only pointing out that "time is counting events" does not work. I believe that I understand my_wan's motivation, I think he wants to avoid the Newtonian idea of time as some undetectable thing which flows and instead focus only on the physical observables. I agree with that goal; I just think that his approach is not right.

That's sounds interesting. Is the undetectable point of view to define time as the proper time along a time-like curve even though no physical clock exists on the curve? Is the detectable point of view to define time via a clock such as the number of vibrations of an atom/light?

 

[Dale]

The undetectable view would be something along the lines of Newton's definition of "absolute" time or Lorentz's definition as "real" time in the aether frame. I wouldn't be bothered with something that is detectable in principle (like the proper time along a curve) even if you don't detect it, but I don't know my_wan's opinion on something like that.

 

[atyy]

The undetectable view would be something along the lines of Newton's definition of "absolute" time or Lorentz's definition as "real" time in the aether frame. I wouldn't be bothered with something that is detectable in principle (like the proper time along a curve) even if you don't detect it, but I don't know my_wan's opinion on something like that.

Why is Newtonian time undetectable in a way that the proper time along a time-like curve isn't?

 

[Dale]

Why is Newtonian time undetectable in a way that the proper time along a time-like curve isn't?

Because Newton defined it that way:

"Absolute, true and mathematical time, of itself, and from its own nature flows equably without regard to anything external, and by another name is called duration: relative, apparent and common time, is some sensible and external (whether accurate or unequable) measure of duration by the means of motion, which is commonly used instead of true time"

So here he expressly differentiates between "apparent" time and "true" time based on the idea that "true" time cannot be detected even in principle.

 

[atyy]

Because Newton defined it that way:

"Absolute, true and mathematical time, of itself, and from its own nature flows equably without regard to anything external, and by another name is called duration: relative, apparent and common time, is some sensible and external (whether accurate or unequable) measure of duration by the means of motion, which is commonly used instead of true time"

So here he expressly differentiates between "apparent" time and "true" time based on the idea that "true" time cannot be detected even in principle.

Oh, I was just thinking that Newtonian time is the t when the equations are expressed in an inertial frame. Would that definition be detectable (in principle)?

 

[Dale]

Oh, I was just thinking that Newtonian time is the t when the equations are expressed in an inertial frame. Would that definition be detectable (in principle)?

I think that would be detectable in principle. I kind of like that definition. It wouldn't be hard to generalize.

 

[Selraybob]

No, the count of a set of spacelike separated events is not measuring anyone's time.

In particular, without a definition of time how do you determine "before" and "after"?

This is where Aristotle got all mucked up with the mystical 'Now'. Now's, according to him, are between the before and after. But before and after are just comparisons of how things are, or how we see them. The car was in the driveway before it was in the street. The clock pointed at 1:00 before it pointed at 1:01. Or backwards. If you stick a Now between the before and after, you have two sets of befores and afters.

Here's what's in my craw though, is that starting with Time and trying to define it doesn't work. The reason Time's a count isn't because that's the great definition of Time. It's because we looked at things changing and then called that Time. It's the changes that are at the center. We're just looking at changes and putting some counter to them so that we can compare other changes to that counter.

It's a paradigm shift. You can even forget little t. You take two things changing and compare them. That's what velocity (or speed, rate, frequency, etc.) is - two counts of two things changing and then comparing the two. Instead of calling the count of the clock little t, you just call it the count of the clock.

And of course, all counters work differently under different conditions. Even the second's defined as a count at a specified temperature.

 

[atyy]

I think that would be detectable in principle. I kind of like that definition. It wouldn't be hard to generalize.

I've read various versions of this definition of Newtonian time ("time is what makes the equations of motion true") in Stephani's and Misner, Thorne and Wheeler's GR texts. That makes it like the proper time and coordinate times of Lorentz inertial frames which exist even in special relativistic field theories which happen not to have any periodic solutions (although in real life theories, we luckily seem to have solutions that are periodic for all practical purposes).

 

[Dale]

The reason Time's a count isn't because that's the great definition of Time. It's because we looked at things changing and then called that Time. It's the changes that are at the center.

Then you should define time in terms of changes, not counts.

I am not going to re-hash the conversation that I just went through with my_wan, go back and see all of the reasons why "time is a count of events" is insufficient. You need to add some constraints, and so far those constraints make the definition circular.

 

[my_wan]

@DaleSpam: @my wan: Is the difference between your points of view that physics is always the study of change dy/dx, but if there is only one "dimension" DaleSpam would call that "space" and my wan would call it "time"?

No that would not work for my opinion and almost certainly would not work for DaleSpam even before reading his response. Neither DaleSpam is making any claims that empirically differ from the standard formulation. I am only arguing that as defined such an event based interpretation works.

 

[my_wan]

No, the count of a set of spacelike separated events is not measuring anyone's time.

You are contradicting yourself saying that "counting events" is completely sufficient and then adding constraints. The "time is counting events" definition is clearly insufficient as you have de facto conceded. You must add such enumerated constraints or clarifications as above.

The causality constraints are particularly problematic for a definition of time. Causality requires a notion of time, so putting a causality constraint into your definition of time makes the definition circular. In particular, without a definition of time how do you determine "before" and "after"?

Btw, I am not opposed to a background independent relational definition of time. I just think that "time is counting events" is way oversimplified. Something this important needs quite a bit more effort and care than that. Many things which qualify as counting events do not qualify as measuring time.

I stated that constraint up fron even before you made the first response to me and in my very first response to you. So no that is not a contradiction.

Consider what it we count in a standard derivation of the Lorentz transform. Take the light mirror where light is bouncing back and forth between two mirrors. We are counting the bounce events at each mirror. If we merely watch the clock we are counting ticks of the clock, etc. In fact we cannot measure "anything" without checking off events that are registered. Whether of not there is a smallest or most fundamental event is immaterial to the argument.

I agree simply saying "time is counting events" is a bit oversimplified. Einstein simply said: "Time is what we measure", and I am saying what we measure are events. The statement really does not go much further than that. It is not fundamentally different. Yet since time effects all else we measure a lot more is involved than just the statement itself. Events can be defined in terms of intervals and that is what SR defines. When we "count" such events we plot them on a graph and curve fit them, taking the limit. But The limiting factors work whether the events involve actual limiting points or not.

 

[GrayGhost]

Why then do physicists persist in using the notion of "time" in explaining relativity? It would be so much easier, more intuitive and correct to talk of (relative) movement and not time.

Can you show how relativity can be defined using relative motion alone in the absence of using space and time?

GrayGhost

 

[Dale]

I agree simply saying "time is counting events" is a bit oversimplified.

Excellent.

Events can be defined in terms of intervals and that is what SR defines.

Yes, I think that is a better approach since the interval is equal to the scaling you were referring to earlier, it contains the causal structure, and it can be used to exclude the timelike sets of events I mentioned.

 

[my_wan]

Excellent.

Yes, I think that is a better approach since the interval is equal to the scaling you were referring to earlier, it contains the causal structure, and it can be used to exclude the timelike sets of events I mentioned.

There are so many ways to look at relativistic effects that are not wrong it is ridiculous. It is only when you try and say what the "real" (absolute) metric of the interval is that it goes over the edge. That makes it a generally a bad idea to think in terms of specific metrics, but the event count works because it does not specify any metric for such events except relativistically.

Whether or not there is a fundamental Planck cutoff or such intervals are maintained in the limit as Einstein imagined is an open question. Since you can just as easily take the limit on events as you can an interval it makes no empirical difference either way. If there is a Planck cutoff it just means your fit become choppy and irregular at very small intervals, macroscopically normalized by a Poisson distribution.

 

[Islam Hassan]

The thing is the thesis of this post: Relativity explained via Movement, not "Time". In fact, we could re-write the title:

'Relativity Explained by L/t, not "t".'

Well enough. I don't think that's too far from what Einstein said, the L/t in question being c, which is a constant in all observers' frames of reference. Sorta QED, no?

In my humble opinion that is a step in the right direction. If I recall correctly, even Einstein himself was initially inclined towards "invariantentheorie" instead of 'relativity' theory. I can't quite recall who prevailed upon him to name it relativity instead.

The benefit to my mind is that by concentrating on the measure of an absolute, invariant movement reference which is c, you avoid saying things like 'time will do this' or 'time will do that', avoid treating time in the agentive form. It is a linguistic point but important to the layman like me who is trying to understand relativity. It is much easier to conceive and understand of material objects which move than immaterial notions that dilate or flow...

IH

 

[Islam Hassan]

Excellent.

Yes, I think that is a better approach since the interval is equal to the scaling you were referring to earlier, it contains the causal structure, and it can be used to exclude the timelike sets of events I mentioned.

Agreed, time as a notion covers both counting and sequencing, the premise being that the sequence is ever-increasing.

 

[Islam Hassan]

Can you show how relativity can be defined using relative motion alone in the absence of using space and time?

GrayGhost

In terms of definition, one needs an absolute movement reference which is c, a set of events and the notion of simultaneity. The movements linked to relative events can be physically measured in space as can c. The instrument which we use to link these is simultaneity which is elaborated in our minds based on our physical perceptions and measurements. The collective set of simultaneous events provides the structure of our notion of time. Time remains a notion and not a physically measured phenomenon.

To redefine relativity in terms of motion alone, I would imagine one needs to benchmark all movement to the simultaneous movement of a photon. To my mind, simultaneity does not imply the physical measure of time, just the physical measurement of co-perceived events. The co-perception is a product of our willed sequencing and measurement, not in my opinion of 'physical' time.

IH

 

[Dale]

There are so many ways to look at relativistic effects that are not wrong it is ridiculous. It is only when you try and say what the "real" (absolute) metric of the interval is that it goes over the edge. That makes it a generally a bad idea to think in terms of specific metrics, but the event count works because it does not specify any metric for such events except relativistically.

I don't understand your point here. All the spacetime metrics that I know of are relativistic. I have no idea what you mean by '"real" (absolute) metric' and why you think that would make it "a generally a bad idea to think in terms of specific metrics" (e.g. Schwarzschild metric).

 

[GrayGhost]

Islam Hassan,

OK. So basically, model relativity w/o the use of time using only the relation of events based upon our sense of simultaneity and using an invariant speed photon as a common reference. IOWs, do what Einstein did, w/o using time, but rather using something else instead. In your estimation, that "something else" should be physical and not a man made notion. For otherwise, you'd simply be replacing what you believe to be one man made notion for another ... and then the big question is, would anything be gained or lost?

Can you give one simple example of how the extent of relative motion would be quantified w/o using time? Have you ever tried this?

GrayGhost

 

[my_wan]

I don't understand your point here. All the spacetime metrics that I know of are relativistic. I have no idea what you mean by '"real" (absolute) metric' and why you think that would make it "a generally a bad idea to think in terms of specific metrics" (e.g. Schwarzschild metric).

The same stuff when you try to ask questions like which clock is "really" go slower. Gr effects makes invariants verses constants issue even more relevant, since it distorts SR in ways that makes the constants more like a choice. Take the apparent mass of an object as its depth in a gravitational field varies. Einstein chose perfectly sensibly to define it as 'apparent' mass. But if you get exactly the same effect if the 'apparent' gravitational constant was what varied, or even the 'apparent' speed of light. In Einstein' words:

In the second place our result shows that, according to the general theory of relativity, the law of the constancy of the velocity of light in vacuo, which constitutes one of two fundamental assumptions in the special theory of relativity and to which we have frequently referred, cannot claim any unlimited validity.

In context this does not invalidate the constancy of the speed of light. This general feature extends to all invariants, not constants. Hence the QG community has their controversy over how much this applies to Planck's constants. This is also the basic theme behind "doubly special relativity". Note why GR gets in trouble with Planck's 'constant' when in terms of Planck length and time:
c = \ell_p/t_p
Planck mass:
m_P=\sqrt{\frac{\hbar c}{G}}
So you can take any first order quantity and express it in terms of pure constants, called the indefinables of physics (space, time, and mass/energy), where all must 'locally' be invariant or more generally some multiple thereof. Yet globally they vary, in the same way light speed does in GR, where the manner in which they vary is both frame dependent and definition dependent, like the apparent mass verses apparent value of G. That is where the so called varying speed of light (VSL) theories get their 'theoretical' legitimacy from. Not from any claim that the speed of light 'actually' changed, as Einstein articulated following the quote I give above.

 

[Dale]

The same stuff when you try to ask questions like which clock is "really" go slower.

Do you have a good definiton of the word "real"? I certainly don't. Without one you will always run into trouble whenever you ask any questions about the reality of anything. That is hardly any kind of argument against that thing.

Gr effects makes invariants verses constants issue even more relevant, since it distorts SR in ways that makes the constants more like a choice. Take the apparent mass of an object as its depth in a gravitational field varies. Einstein chose perfectly sensibly to define it as 'apparent' mass. But if you get exactly the same effect if the 'apparent' gravitational constant was what varied, or even the 'apparent' speed of light. In Einstein' words:



In context this does not invalidate the constancy of the speed of light. This general feature extends to all invariants, not constants. Hence the QG community has their controversy over how much this applies to Planck's constants. This is also the basic theme behind "doubly special relativity". Note why GR gets in trouble with Planck's 'constant' when in terms of Planck length and time:
c = \ell_p/t_p
Planck mass:
m_P=\sqrt{\frac{\hbar c}{G}}
So you can take any first order quantity and express it in terms of pure constants, called the indefinables of physics (space, time, and mass/energy), where all must 'locally' be invariant or more generally some multiple thereof. Yet globally they vary, in the same way light speed does in GR, where the manner in which they vary is both frame dependent and definition dependent, like the apparent mass verses apparent value of G. That is where the so called varying speed of light (VSL) theories get their 'theoretical' legitimacy from. Not from any claim that the speed of light 'actually' changed, as Einstein articulated following the quote I give above.

None of this seems to indicate why using the metric is a bad idea.

As you and I have already agreed, you cannot simply count events to get time, you must add some additinal constraints. In particular you mentioned causality and scaling, both of which are provided by the metric. So I don't see any difference between defining time in terms of counting events with scaling and causality constraints and defining time in terms of the metric.

 

[Islam Hassan]

In your estimation, that "something else" should be physical and not a man made notion. For otherwise, you'd simply be replacing what you believe to be one man made notion for another ... and then the big question is, would anything be gained or lost?

GrayGhost,

Yes I believe something would be gained, at least for laymen like myself. The phenomenon which led me to consider movement instead of time as the focus of an explanation for relativity is that of 'time' stopping at the speed of light (or at anything approaching the speed of light for a massive object). I found this very difficult to understand intuitively. Then I tried to reason from a movement perspective and found that an intuitive explanation suddenly snapped into place. The following is how it happened:

- A fan say, or any mechanical object traveling at 0.99999... the speed of light would gain mass according to special relativity. The closer it gets to c, the more massive it becomes and as it tends more to c its mass tends to infinity;
- It takes infinite energy to move a mass tending to infinity; and
- Therefore with the limited electrical energy feeding its motor, the fan's movement will tend to absolute immobility.

Explained this way, the layman that I am understands the phenomenon easily and intuitively. Explained in terms of the semantically funky 'time dilation', such is not the case at all.

I can quite understand that seasoned physicists are quite used to the present terminology and don't really need any other 'interpretation' to make sense of SR; my focus, as I said is on the layman for whom relativity can be quite a daunting proposition...

This is one example but I am sure there are others. My basic thesis therefore is that explaining relativity at the most basic, phenomenological level makes it more easily comprehensible. Just a lay opinion of course, but there you go...

I'm still working on your challenge re a simple example of how the extent of relative motion would be quantified w/o using time; get back to you soon.

IH

 

[dm4b]

All the spacetime metrics that I know of are relativistic.

Euclidean Metric

 

[my_wan]

Do you have a good definiton of the word "real"? I certainly don't. Without one you will always run into trouble whenever you ask any questions about the reality of anything. That is hardly any kind of argument against that thing.

As far as the word real, if someone is asking which clock is really going slower they are almost certainly assuming a Euclidean metric, such as in Newtonian physics. We know the absolutes this entails to be wrong. Hence specifically what is 'meant' by the the word 'real' is moot when the question itself entails demonstrably invalid assumptions.

None of this seems to indicate why using the metric is a bad idea.

Which metric? A Euclidean metric is demonstrably wrong in the general case. In terms of scale any coordinate choice is a nonphysical choice even if you apply the transforms needed to give a Euclidean metric some level of observer dependent validity. By assuming time is an independent variable from the events it 'smells' of Newtonian time with an added feature of malleability. This is not so.

Suppose you go 86% the speed of light getting a dilation factor of 1/2. To you time did not dilate, the distance to your destination was cut in half. Now suppose you decrease gravitational depth to the field around the ship to erase the relative time dilation factor between you and back home. Only now, without changing your proper motion from Earths perspective, you see as measured by you that you have twice as far to travel to get to your destination, hence will take you twice as long. The same amount of time it appears to take you from Earths perspective.

You cannot separate time from the events you observe under any circumstances, at least not without playing 'what if' games with energy conditions outside observable reach.

As you and I have already agreed, you cannot simply count events to get time, you must add some additinal constraints.

I agreed that saying time is a "bit oversimplified", not invalid. I would have left the debate there except that you quoted that naked sentence while the qualifiers that followed was ignored. In fact you can (and it is in fact what we do) to not only define time but to derive all the relativistic Lorentz transforms when event counts do not match.

When we measure something we record a series of events. Then we curve fit these events as if they occur on a continuum. Whether they actually do exist on a continuum or not is immaterial, as what we empirically record is not the continuum but a discrete series of events. So in fact everything we empirically know about time was derived from a discrete series of events, i.e., "what we measure".

In particular you mentioned causality and scaling, both of which are provided by the metric.

So you are using metric in the very narrow sense of a particular type of metric. Euclidean metrics explicitly require transforms for the very fact that they are not valid in the more general case. You say they are "provided by the metric", implicitly meaning Lorentz metric rather than metric. Even so you have a major reversal of logic here. The event count is not "provided by the (Lorentz) metric", the Lorentz metric is provided by the event count. Such as flashes of light on a train in Einstein's SR derivation. This reversal of logic is confusing to a lot of people. The empirical events justify the theory, the theory does not justify the empirical data.

So I don't see any difference between defining time in terms of counting events with scaling and causality constraints and defining time in terms of the metric.

Again, you implicitly referred to the Lorentz metric 'only' as if it was "the" metric. Yet the Lorentz metric was explicitly derived from discontinuities in counting the same discrete events from alternate perspectives and noting that they did not match the local discrete event counts in each of those perspectives. If no events occur there is nothing to count or do the counting hence the notion of time is moot. It is in fact discrete event from which the Lorentz metric was derived, however much we theoretically idealize it after the fact.

 

[atyy]

It is in fact discrete event from which the Lorentz metric was derived, however much we theoretically idealize it after the fact.

But if you want to have "theory-free" definitions, how can you even define an event?

Also, if you didn't need the theory to define time, how do you explain that we have theories with a preferred time, as well as theories with many times?

 

[Dale]

Which metric? ...
So you are using metric in the very narrow sense of a particular type of metric.

This is a relativity forum, so I am using "the metric" in the usual sense in relativity. I.e. the unique metric for a given spacetime manifold. There is only one for any given spacetime, so it is not just narrow, it is unique. It is never Euclidean so the bulk of your post is rather irrelevant.

I thought that this standard meaning of the phrase "the metric" would be well-understood by anyone posting here; I apologize for my poor communication.

The event count is not "provided by the (Lorentz) metric"

I never said it was. The scaling and causality constraints are provided by the metric. When you take a set of events, add a causality constraint, and scale it appropriately, and call the result "time", then you have done nothing more nor less than evaluate the metric between the events.

 

[my_wan]

But if you want to have "theory-free" definitions, how can you even define an event?

Also, if you didn't need the theory to define time, how do you explain that we have theories with a preferred time, as well as theories with many times?

I am not talking about "theory-free" definitions. I am talking about the fact that theory is empirically driven, not the other way around.

With respect to "preferred time" do you mean direction? I do not want to answer that presumptuously as I can relate it to too many concepts. However, I do not know in what way I said you do not need "theory to define time". I said that the legitimacy of theory is predicated on and justified by the empirical data, not the other way around. Without which it is nothing more that philosophy. And that this empirical data, as it relates to time, was in fact the result of counting discrete events. The fact that we need a theory for operational consistency does not raise theory above the empirical constraints it is predicated on.

My initial statement, which has been taken issue with, is that time can consistently be viewed in terms of event counts. A bit simple, but fully defensible empirically. When Einstein said "time is what we measure" it was simplistic also. My statement merely boils down to defining "what we measure", and that is discrete series of events. Theoretically it works whether it comes in discrete steps or not, but the fact remain: We measure discrete series of events.

 

[Dale]

My initial statement, which has been taken issue with, is that time can consistently be viewed in terms of event counts. A bit simple, but fully defensible empirically.

The complexities glossed over by the "bit simple" are important, as we have discussed. An arbitrary event count is NOT a measure of time without the causality and scaling constraints.

 

[my_wan]

This is a relativity forum, so I am using "the metric" in the usual sense in relativity. I.e. the unique metric for a given spacetime manifold. There is only one for any given spacetime, so it is not just narrow, it is unique. It is never Euclidean so the bulk of your post is rather irrelevant.

I thought that this standard meaning of the phrase "the metric" would be well-understood by anyone posting here; I apologize for my poor communication.

Are you saying that because I am on a Relativity forum I cannot use "metric" in the general sense, which includes any choice of coordinate system? It is GR that prides itself the most on treating all coordinate systems equally.

Perhaps some of these points are a bit on the pedantic side, but no more or less so than taking issue with defining time in terms of event sequences.

I never said it was. The scaling and causality constraints are provided by the metric. When you take a set of events, add a causality constraint, and scale it appropriately, and call the result "time", then you have done nothing more nor less than evaluate the metric between the events.

Funny, I can turn this around again and say "causality constraints" is not "provided by the metric", the "metric" is provided by the "causality constraints".

I wonder, how much of this inversion of logic is behind many of the established controversies in science?

 

[my_wan]

The complexities glossed over by the "bit simple" are important, as we have discussed. An arbitrary event count is NOT a measure of time without the causality and scaling constraints.

The same constraints I qualified with both before you responded and in my first response to you. The "bit simple" was no more or less "bit simple" than "time is what we measure". That was the point of the post where I said that. Only you chose to chop out the "bit simple" part for your quote of me.