# Teaching Relativity in a College Physics course

1. Dec 5, 2015

### Mister T

I teach a bit of special relativity to non-majors enrolled in the typical two-semester introductory college physics sequence. These are my goals.

1. Have the students develop an understanding of the basics such as length contraction, time dilation, relativity of simultaneity, energy, and momentum.

2. Spend only one week of class time.

You may immediately conclude that both goals cannot be acheived. I'm willing to budge a bit on either one, but if possible I'd like to go as far as I can towards acheiving both.

I have only two meetings during this week, so I start with a reading assignment and an online homework assignment, both of which are to be completed before the first meeting. The reading assignment is half of the chapter on relativity found in the typical College Physics textbook. (The other half of the chapter is the reading assignment for the week's second meeting). The online homework assignment consists of numerical calculations involving the relationships between $v$, $\beta$, and $\gamma$.
$$\beta=\frac{v}{c}$$
$$\gamma=\frac{1}{\sqrt{1-\beta^2}}$$
During the first meeting I go over the stuff I discussed above in Goal 1. In the past I've not been happy with the outcome and I'm thinking that a bit of spacetime geometry may help. The reason I think that is because the geometry is another representation, and multiple representations lead, in my belief, to a deeper understanding. Or they at least offer an alternative way to understand. Anyone who's studied relativity knows how valuable spacetime diagrams are, and the textbooks I've used do not take advantage of that.

So, with that in mind I've written up what I plan to be the lesson for that second class meeting. It's the attached PDF. If you take the time to read it and have feedback I'd appreciate hearing from you. Thanks.

File size:
216.6 KB
Views:
192
2. Dec 5, 2015

### Andrew Mason

You do not appear to explain or even mention the fundamental principles/postulates of SR. You also do not deal much with inertial frames of reference. I am not sure why these are omitted. It is impossible to understand anything about SR if you don't explain that the speed of light is the same in all inertial frames of reference and that all inertial frames are equivalent i.e. the laws of physics are the same in to all inertial observers. I don't know how you expect anyone to understand the proper time, time dilation and length contraction without first explaining the Lorentz transformation.

If I were teaching this I would start by explaining the results of Michelson Morley experiment and compare the approach of Lorentz and Fitzgerald to that of Einstein. Then I would explain how Einstein's approach - abandoning the concept of the ether - explains the Michelson Morley results. Then I would introduce space-time diagrams and the Lorentz transformations. This allows one to calculate time dilation and length contractions. If you can do that in one week, good luck!

AM

Last edited: Dec 5, 2015
3. Dec 5, 2015

### Mister T

Much of that is addressed in both the reading assignment and the first lecture. That's the reason you don't see it in the attachment.

4. Dec 5, 2015

### nrqed

Good job. But I do not understand the nee to introduce units with c=1 in that context. I don't see the benefit and I see a lot of possible confusion. Those units are just useful for research scientists who prefer not to have to remember where to put factors of c but it is overkill and just confusing for first time students. I would get rid of that and include all factors of c.

Also, in the muon example I think it is absolutely necessary to emphasize the fact that $\Delta x'=0$ (with primed coordinates attached to the muon). My students are always puzzled at first, even though it is obvious to us.

But like the other poster said, I think it is important to show that the invariance of the spacetime interval follows from imposing that the speed of light is the same in all frames.

5. Dec 5, 2015

### Andrew Mason

It might help if you provided us with your plan for the first lecture. It is very odd that you don't relate anything in the second lecture to the first. Frankly, I found your approach very unclear and confusing. Your explanation of speed as minutes/minute and time in minutes is very odd and confusing. I would also suggest using space-time diagrams to show the way the axes (of distance and time in other inertial frames) are transformed using the Lorentz transformations.

AM

6. Dec 6, 2015

### vela

Staff Emeritus
I suggest you look at chapter 1 of Spacetime Physics by Taylor and Wheeler. I just covered relativity in my class trying a spacetime geometry approach, and there's a lot of blanks to fill in for the students. The students also have their textbook, which uses the more typical approach that doesn't really mention geometry, so they got to see the theory from two perspectives.

I don't have a problem with $c=1$ for the geometrical approach. Typically, if the Lorentz transformations are mentioned at all in the usual textbooks, they are written as
\begin{align*}
t' &= \gamma(t - \frac{vx}{c^2}) \\
x' &= \gamma(x - vt)
\end{align*} which makes it look like $t$ and $x$ are on different footing. In contrast, when written as
\begin{align*}
t' &= \gamma(t - \beta x) \\
x' &= \gamma(x - \beta t)
\end{align*} where $t$ is measured in units of length, the symmetry between $x$ and $t$ is readily apparent. It's not terribly difficult for students to understand that when you say t = 1 m, you're talking about the time that it takes light to travel 1 meter. It's really no different than when someone says "Disneylan is an hour away." People know that an hour isn't literally a distance, but they understand it's the distance that corresponds to traveling for that amount of time.

7. Dec 6, 2015

### nrqed

I still do not see the advantage, at the introductory level. There are enough subtleties to grasp without adding an extra layer which does not serve much purpose, it seems to me (and time and space really are on different footing, as the difference of sign in the metric makes clear).

In any case, even setting c=1 in the equations does not dispense totally from needing to use c. If the muon's lifetime is give as 2.2 microseconds, the students will then have to covert that to a distance using c before plugging the result in the equations. If I had only one work or so to teach relativity, I would not want to spend some time discussing this layer of complication that does not bring in much insight, I think (and I know that from *our* point of view it is easy and useful, but from *their* point of view, I don't see the advantages.) To me, it is more pedagogically sound to try to introduce as few new concepts a at a time, especially when they are extremely counterintuitive. I think it is challenging enough to explain relativity or simultaneity, time dilation, etc without adding what seems an unnecessary layer at that level.

Just my opinion, of course.

8. Dec 6, 2015

### Mister T

The first post gives an overview of what's covered not only in the first lecture, but more importantly what the students are exposed to in the way of the reading assignment and the homework assignment. You can look, for example, at the OpenStax College Physics book by downloading the PDF for free.

One of the things covered there is time dilation, and that is referred to in the second lecture. Another is the relation between $\beta$ and $\gamma$.

The second lecture focuses on proper time and time dilation.

9. Dec 6, 2015

### Mister T

Thanks. The reason for using units where c = 1 is so that the geometry is clear. Otherwise it's obscured. I believe that's the reason Vela mentioned Taylor and Wheeler's book. It's not clear from the opening paragraph in my attachment, but I plan to spend time in class expanding on the notion that the invariance of the length of line segment in 2-d Euclidean space would be obscured if each dimension were measured in different units.

Yes, I spent some time thinking about that. I decided I'd see if I could get by without ever introducing x' and t'. It may not work, but if I want to meet Goal 2 (spend one week) I have to sacrifice something. I won't introduce invariance of the interval except insofar as to use proper time as the length of timelike intervals. Spacelike intervals are another thing that I sacrifice.

I will definitely stress in both lectures that everything is a consequence of the two postulates. That is also something stressed in the reading assignments in the textbook, too.

My philosophy of teaching is to not necessarily be the first presenter of everything. Some things are presented to the student in reading assignments and homework assignments. I ask myself what is the best way to have a topic presented, and make my choice of methods based on that.

10. Dec 6, 2015

### Mister T

That's what I'm wanting to try, too. If I can get students to understand time dilation and the twin paradox I'll be happy. Well, happier, anyway.

I don't know if you're familiar with Paul Hewitt's old cartoon video called Time Dilation. It's really well done, and makes use of the Doppler effect to demonstrate the twin paradox. It matches the treatment in his textbook. The numbers in my lesson match the ones Hewitt uses, so the students will see both treatments (Doppler and spacetime geometry).

Last edited: Dec 6, 2015
11. Dec 7, 2015

### Andrew Mason

My concern about this approach is that in a one-week introduction to SR you are going to leave students confused unless you relate the space-time geometry to the underlying physics. The essential element is the constancy of the speed of light for all inertial observers.

Consider the Einstein light train thought experiment where a light signal from the origin occurs when the tail of the moving train coincides with the origin in the rest frame at time t=0 for observers A and A' in the respective frames. There is a mirror at the nose of the train and mirror at x = L in the rest frame, where L is the proper length of the train. A measures the time the signal reflected from his stationary mirror at x=L reached him as $t_0 = 2L/c$. The A' observer measures the same time for the signal to reach him: $t_{train} = 2L/c$.

But to the stationary observer, it took longer than L/c to reach the mirror and less than L/c to reflect back to A'. This is because A sees the light traveling farther than L in order to reach the mirror at the nose of the train and shorter than L to return. Working out the actual time difference is not trivial. It involves derivation of the Lorentz transformation. Einstein took several pages to show this in his 1905 paper.

The reason $(\Delta t_0)^2 = (\Delta t)^2 + (\Delta x)^2$ (as you choose to write it with units of x being units of time rather than distance) is not easy to see. Yet you make it look so simple that the students are going to think they are stupid if they don't see it. And that is not a good thing from a pedagogical point of view.

But the reason there must be a time difference is not difficult to see. And that is probably what the students are going to remember about special relativity.

AM

Last edited: Dec 7, 2015
12. Dec 7, 2015

### nrqed

But all terms have the same units, of course, even if we keep c not equal to 1. Then there is $(c \Delta t)^2$ in the square of the spacetime interval which clearly has the same units as $(\Delta x)^2$. I still don't see the advantage of having c=1. With that convention, two things happen: A) when students see $(\Delta x)^2 + (\Delta t)^2$, they have to remember that this weird system of units is used in order to see that the two terms have the same dimension and B) every time they have a time given in a problem, they cannot leave it in seconds in the equations, they first have to convert it to meters before being able to use the equations. For people who will become physicists, the c=1 system is useful because it simplifies equations but I still don't see at all how this can possibly help people who will not do a career in physics understand better relativity.

I personally feel that it is absolutely necessary to explain that $\Delta x' =0$. I have had students learning the time dilation or length contraction formula without understanding that these formula are not always valid! Then, whenever they see a $\Delta t'$ given in a problem, they assume that $\Delta t = \gamma \Delta t'$ which if of course not always correct. And when this is pointed out to them, they get all confused if they haven't learned the proper use of the equations the first time around.
If you introduce the concept of spacetime interval, why not calculate it in S' and then use the fact that it must be the same in S? It seems to me that all the work that went into introducing the spacetime interval is lost if it is not fully used in all problems.

But again, that's just my humble opinion.

13. Dec 8, 2015

### Orodruin

Staff Emeritus
The sad part here is that even in higher relativity courses, students get this wrong because they are used to it always applying in the problems they have seen so far in earlier courses. Even if you are discussing the derivation of the length contraction formula in detail and warn about its use for objects which are not moving at constant velocity, you can safely bet on that someone (most) will try to apply it when you ask them to figure out what is going on in Bell's spaceship paradox.

The $\Delta x' = 0$ part is really difficult to get out of students preconceptions. In particular, this is not made easier by some textbooks in modern physics making statements such as "the muon travels 600 m in its rest frame" ...

14. Dec 8, 2015

### nrqed

Yes, this last quote is indeed terrible!

I always do a few examples where neither of the $\Delta x, \Delta x', \Delta t, \Delta t'$ are zero (for example, Joe throws a ball to Jack, with both of them in S', and I have y students find the time and space intervals in S, and make them check that the ball's velocity in S is indeed given by the relativistic velocity addition law. I think it is not pedagogically sound to only do examples where at least one of the intervals is zero as it gives the impression that some special formula are always valid (like $\Delta t = \gamma \Delta t'$).

15. Dec 11, 2015

### Mister T

Thanks to all for taking the time to contribute. I really appreciate it, and I'm still actively working on this project. Finals week and its approach have kept me busy with less desirable tasks (I love teaching, hate grading, but you can't have one without the other).

That's a very good point. The textbook reading assignments and my first lecture specify how the 2nd Postulate fits into the topics of Goal 1, but I would like to find a way to fold it into the lesson on spacetime geometry. Any suggestions on how to do this given the constraints of Goal 2?

I'm sure you meant to write $(\Delta t_0)^2 = (\Delta t)^2 - (\Delta x)^2$. Your point is very well put. I need to address this issue. But how to do it? I'm thinking ...

I do not plan to introduce a primed coordinate system. I realize the importance of stressing that $\Delta x'=0$ is a necessary condition for $\Delta t'$ to be a proper time. But the point has been well made in this thread that the only way to address that issue is to effectively communicate situations where there is no proper time in either of the frames being considered (for a timelike separation of two events). The way I'm doing it is to simply stress that the reason it's a proper time in this particular frame is because the one clock being used to measure the time interval is present at both events.

The vast majority of students taking introductory physics in colleges in the US are enrolled in a sequence of courses that are labelled either "calculus-based" or "trigonometry-based". The lesson being discussed here is for the latter. I have been reviewing textbooks used in both sequences and almost all of them never consider two frames where there is neither a proper time nor a proper length in one of them. The section in the chapter devoted to "relativistic addition of velocities" is the only situation that's the exception, and even then they don't explicitly refer to a primed coordinate system.

16. Dec 11, 2015

### Mister T

Every time I have taught the course I've regretted not teaching the students about the simplicity of a system where c=1. There are many encounters in this course where it's an advantage to analyze a situation using units like light years and years. Almost every textbook I looked at works at least one example this way.

N. David Mermin uses the phoot and the nanosecond. (The phoot is the distance light travels in a nanosecond, equal to 0.299 792 458 m, which is less than 2% different from the foot, which is 0.3048 m). As has already been mentioned in this thread Taylor and Wheeler use meters of time.

I do agree that the discussion of minutes of distance in my lesson may be confusing. I'll see what I can do to clean it up. The issue I ran into when writing it is the following. I want students to more easily be able to switch back and forth between units where c=1 and SI units. One way to do that is to use $\beta$ instead of $v$ in the equations. But if you do that you can't use, for example, years and light years. $v$ would be measured in light years per year, $c$ would be 1 light year per year, and you have to use $v=\beta c$ to get the units to come out right. Maybe this concern is overly pedantic, and if I can convince myself of that it may help me write that discussion of minutes of distance in a way that's less confusing.

17. Dec 12, 2015

### nrqed

I also think that working in light years (and in years) is good for many problems (when talking about traveling to other solar systems for example). But I still don't see the need for setting c=1 (sorry if I sound like a broken record). All I do is to tell my students that by definition, 1 ly is the distance travelled at the speed of light in one year, so that although we could in principle convert it to meters, it is much simpler to write instead 1 ly = 1 year x c . So, for example if a ratio of a distance over the speed of light enters a calculation, we directly get that 1 ly/c = 1 yearc/c = 1 year. If we divide, say, 20 ly/40 years, we directly get half the speed of light, c/2.

Everything works well and the units confirm that the calculation was done correctly, so the presence of units is helpful as a mean to reassure the students that their manipulations are probably correct. We don't care about units normally but when encountering these new strange formula with weird physical results, I think that having the units shown explicitly is an added help.

I will stop arguing my point since I am the only one thinking this way but one final thought: with c=1, sometimes results that have no dimensions are really meant to be a speed whereas other numbers that are dimensionless are truly dimensionless. I know that for us, the teachers, it is trivial to tell when a factor of 1/2 somewhere truly means 1/2 c and a factor of 1/2 somewhere else truly means 1/2, we just see what is supposed to be a speed and what is supposed to be a pure number. I don't think that students should have to even think about this while they are struggling to understand something as counterintuitive as SR, but that's just me.

18. Dec 12, 2015

### Orodruin

Staff Emeritus
You do not need to set c = 1, but it highlights the structure and symmetry of the theory. Furthermore, as space-time is a manifold, it is somewhat artificial to measure different directions in different units, in particular when doing coordinate transformations. Imagine living on the sea. There are then two directions you probably wish to measure in relatively large units, such as kilometers, and one direction you will probably use smaller units for. This is of course not wrong, but it does not highlight the fact that the spatial directions are equivalent (simply because for you the presence of the sea seemingly breaks the equivalence). Let us say we measure lengths parallel to the surface in km and depth in feet (to underline the fact that these units a priori are independent). The trajectory along which something moves underwater then has a rate of descent measured in the unit 1 foot/km which is a universal constant. With the introduction of Minkowski space-time which underlines the symmetry between space and time, c is very similar to the 1 foot/km in the above example. Sure, it may be convenient to work in such units, but it is somewhat artificial.

19. Dec 12, 2015

### Mister T

I appreciate you highlighting the issue. So you see c = 1 to be different from c = 1 ly/y, which of course they are. This is the issue I referred to above that I was thinking about when writing the lesson.

This is analogous to saying the slope of a line in Euclidean 2-space is a dimensionless number. If you were in the habit of using a system where x and y were measured in different units you'd run into the same issue. The foundation of spacetime geometry is that the relationship between space and time is a slope.

20. Dec 12, 2015

### nrqed

But I think of the coordinates as being $\Delta x$ and $c \Delta t$, which have the same units. It is because you insist on using $\Delta x$ and $\Delta t$ that you find yourself forces to introduce strange units (from the point of view of students) with c=1. But I won't insist :-). Good luck with your classes, it is a challenging topic to teach!

21. Dec 12, 2015

### Andrew Mason

I don't follow this. c can't be equal to a dimensionless "1". It has units of distance/time. So if the magnitude of c is equal to 1, it is implicit that the unit of distance must be equal to the distance that light travels in a unit of time. So c = 1 ly/y or c = 1 phoot/nanosecond.

This suggests that time and space are physically the same. They aren't. Time is measured in units of time and space is measured in units of distance.

I agree.

AM

Last edited: Dec 12, 2015
22. Dec 12, 2015

### Orodruin

Staff Emeritus
Which they are! This is a very fundamental insight in relativity! One observer's time direction is a time and space direction (although time-like) for another!

23. Dec 12, 2015

### nrqed

Then what do we get if we add 5 minutes to 2 meters? :-)

I would rather say that a fundamental insight of relativity is that the speed of light is an invariant, which then implies that $c\Delta t$ is closely related to $\Delta x$, and therefore a value of $c \Delta t$ in one frame shows up as a combination of $c \Delta t'$ and $\Delta x'$ in another frame. I really don't see why it is necessary to have c=1 to be able to appreciate all this. I understand that it simplifies calculations but setting c=1 is only a special choice of units (and it is actually cheating to say that c=1 since what we really mean is that we choose units of distance "uod" and of time "uot" such that c= 1uod/uot and we lazily drop the units). And a choice of units has absolutely no fundamental meaning!! I think that here it just obscures the essential role of c in making it disappear from the equations.

24. Dec 12, 2015

### Orodruin

Staff Emeritus
You would not add something so small as 5 m to 2 s. One of them is about 8 orders of magnitude larger than the other ...
The introduction of c only acts to obscure the symmetry of the theory and the fact that space and time are really only two sides of the same coin.
No, this is just plain wrong. It is as valid to consider units of time and units of distance to be equivalent, just as units of depth and units along the ocean surface. There is absolutely no difference here. In both cases, the directions are related by a coordinate transformation and all you are doing when introducing c is using different units for different coordinates. This does not affect the manifold itself. It has absolutely nothing to do with laziness and everything to do with treating space and time on an equal footing - which should be done because they are really only different directions on the same manifold. It is not only that what shows up as time for one observer shows up as time and space for another. It is that they are fundamentally intertwined and that there is absolutely no way of separating them in a physically unambiguous way.

Regardless of what you think of it, this is the preferred modus operandi for the large majority of physicists. If you dislike the use of t when doing this, introduce the new time coordinate $x^0 = ct$. The c does not even help you in dimensional analysis as it always comes along with the t (or dt). The only thing you can get out of it is whether or not you introduced enough powers of c, which is irrelevant if c = 1.

25. Dec 12, 2015

### Andrew Mason

The fundamental insight into relativity is that space and time are related. While an interval of time in one inertial reference frame may appear to be an interval of both space and time in other inertial frames of reference, space and time are fundamentally different physical concepts. An interval of time in one inertial reference frame cannot be seen as an interval of space in another. An interval of space (i.e distance between two simultaneous events) measured by one observer cannot be seen as an interval of time by another.

This same kind of confusion arises when talking about mass and energy 'equivalence'. Mass and energy are very different concepts. Mass may be related to energy content of a body, but mass and energy are very different concepts. with very different units and very different physical attributes.

AM