Physics vs. Chemistry: Nobody Does Research on Newtonian Mechanics

Not long ago, I had a meeting with the Dean, who is a chemist. One of the things I talked about was my plan for distributing teaching assignments in the next few years, which ran into an interesting cultural difference. I explained how I was trying to make the distribution of assignments a little more regular and uniform, getting everybody to teach both intro and upper-level courses, and he said (paraphrased), "That's funny. We never have a problem with that in chemistry-- the organic chemists teach Orgo, and the rest of us teach general chemistry, and that's that."

It took me a minute to put my finger on what the difference was, but eventually I got it. The problem is that while introductory organic chemistry is pretty directly related to research in organic chemistry, the same isn't true for physics. We all have to know classical physics, but basically nobody studies it for their research (with a partial exception for people doing physics education research, but they're mostly researching more effective ways to teach intro mechanics). The physics that we use in research is qualitatively different than what we teach in class-- there are a few places where you can construct relevant examples, such as using magnetic fields to distinguish charges of subatomic particles, but most actual physics research is way beyond the intro curriculum.

This creates a real difference when it comes to assigning courses. There's no obvious reason to assign anybody in particular to do intro classical mechanics every year, or intro E&M, because they're both more or less equally distant from everybody's research topic. The stuff that's actually relevant to our research doesn't start until sophomore level modern physics, but we only offer one section of that a year. And the really research-relevant material doesn't make it into the classroom until the "Special Topics" courses at the 300 level, which we rotate through on a three-year cycle.

This is basically just another problem caused by the need to teach "old stuff" in physics classes. And as with all the other issues there, it's hard to see what can be done to improve the situation. For a variety of reasons, we're constrained to teach introductory classes that don't connect all that well to our research, mostly in service of students in other majors who need to know old physics for one reason or another. That makes it harder to find a "fit" for introductory teaching, and increases the demand for teaching the very small set of upper-level courses that are more relevant to research, which thus requires more of a plan for assigning those courses. I suspect that Physics is more like Math than Chemistry, in this respect: the mathematicians have a similar need to rotate everyone through the intro courses, because nobody does research on introductory calculus.

I don't have any really deep point here, it just struck me as an interesting bit of academic culture. Though if anybody has any brilliant ideas on how to tie introductory Newtonian mechanics to current research in a more direct way, I'd love to hear them.

More like this

Typo or quantum mechanics?

"The physics that we use in research is qualitatively different than what we use in research"

By Rob Aiello (not verified) on 14 Feb 2013 #permalink

Hitting "Publish" rather than "Save" when a student came in with a question. Fixed now, though I was tempted to leave it alone.

I would say that the people who do research on classical mechanics are the nonlinear dynamics and chaos types, and certain types of astronomers. However, there aren't enough of these people out there to assign all freshman mechanics courses to them.

(Somebody might say something about physicists using Lagrangians. They are usually studying fields, not discrete systems that get studied in a course on Lagrangian mechanics. And none of that shows up in freshman Newtonian mechanics.)

Plenty of physicists use E&M in their research, but few study E&M as E&M. They take it as a given and apply it, for the most part.

There are people who actually study quantum mechanics rather than just using it. i.e. Their question is actually "How does quantum mechanics work?" rather than "How does this system behave? I shall use the Schrodinger equation to find out!" And although they get less attention, there are people who actually study statistical mechanics as its own subject rather than just using it as a tool to calculate the properties of whatever systems they're really interested in.

I know a few people who use classical mechanics in their research, but they're mostly astronomers who study orbital dynamics and either have perturbation equations to basically deal with more than two body interaction and/or have problems that are so full of math that the computer has to keep track of a million planetesimals orbiting a star*. My first paper was on the rotational motion of a funny-looking moon under quickly-varying torques now that we had new data on it.

But it's not so much as directly studying the physics as studying how to apply the physics, like how many scientists use calculus all the time, but they aren't advancing calculus, just its application.

I don't know enough about chemistry to compare what I learned in my general chem courses versus modern research: some of it, like gas laws and such, are probably in the same boat. They're understood and used, but no one studies them, just applies them.

* And a lot of times it's not strictly classical: one of those perturbations is 'oh, yeah, and there's relativity' or the catch-all 'non-gravitational forces' (which usually means magnetic fields or radiation are involved).

By Rebecca Harbison (not verified) on 14 Feb 2013 #permalink

A significant amount of plasma physics involves looking for non-vacuum solutions to Maxwell's equations. Allowing ρ and j to be nonzero makes things get interesting in a hurry, particularly if you also allow finite temperatures and a background magnetic field.

As for people working with classical mechanics, there is fluid dynamics, as well as the orbital dynamics other posters have mentioned. Fluid dynamics also gets interesting. Think you can ignore viscosity because it's so much smaller than inertial forces? Sorry, there's an additional boundary condition attached to the viscosity term in the Navier-Stokes equation. You get turbulence instead.

You can even combine fluid dynamics with plasma physics, getting magnetohydrodynamics.

But all of these areas are somewhat specialized, and not all departments (particularly small ones) will have people working in these areas. So your rotation system makes more sense in your case, especially since the introductory classes tend to be more work due to their size--physical science and engineering majors, as well as physics majors, have to take them.

By Eric Lund (not verified) on 14 Feb 2013 #permalink

Alex pretty much has my comment, but dude: solitons. Solitons solitons solitons. I can explain pretty much everything I'm doing in freshman physics terminology, you just have to accept that there's going to be some nasty nonlinearities. If you feel the need to be quantum go stick a soliton in your Bose-Einstein Condensate and call yourself Randy Hulet.

I'd also say that there's plenty of interesting continuum mechanics research going on (including in fluid dynamics), even if it's happen more in your aeronautics or mechanical engineering departments.

Oh, and granular mechanics? All classical. And crazy as hell.

I think part of the issue comes down to what (if any) is the distinction between research "in" Newtonian mechanics vs. research that "merely uses" Newtonian mechanics.

But how does that differ from, say, general chemistry or basic biology? No-one's actively reproving mass balances or basic atomic orbital theory.

And how close, exactly, is sophomore orgo to the kind going on in research labs?

And after this thought, perhaps an interesting point: Should we ask our engineering faculties (maybe get a prof from MechE and one from EE) to be the ones teaching basic physics? They're the ones working with it every day.

I think the using vs researching issue is key. A lot of people are trying to use the Classical Physics models in new situations, but hardly anybody is really doing work on improving or testing the classical models now.

On the Quantum side a lot of people are still testing the model itself, trying to create experiments where if predictions are wrong, the model is suspect more than our understanding of the circumstances of the experiment. From the high energy Higgs searches to attempts to get macroscopic mixed states there is a lot of activity trying to find where the model breaks.

Proposed for discussion: physics for physics students shouldn't waste time on classical except as a small-angle approximation.

For those non-physics students who need to get into the classical weeds ... that's what engineers do. The big problem there is more political and budgetary than curricular.

By D. C. Sessions (not verified) on 14 Feb 2013 #permalink

How many physics professors are actively investigating the foundations of QM or relativity? I realize many in the AMO community are more concerned with this (oh look, we built the smallest two slit experiment...), but look at the folks doing biophysics, condensed matter, astronomy. Your hands are more than full with things we've known about for years.

If I look at much of the research my friends in physics are doing, it tends to be much more "applications"--build this thermoelectric material and characterize its phonon spectrum, develop a new type of microwave sensing antenna, explore these geodynamical interactions, develop a new nanoscale waveguide. I'd argue that most physics research is not the kind of foundational stuff that Aspect, Kimble, etc. are known for but instead looking at the consequences of these foundations.


I agree. Given that most physicists are exploring the consequences of these foundations, there are two interesting questions to ask:
1) What is the difference between a physicist exploring the phonon spectrum of a material (or whatever) and a materials scientist, engineer, or whatever, exploring the phonon spectrum of that material? Is there a difference?

2) If there is a difference, does that difference make the physicist better-suited to teach freshman mechanics and E&M, as opposed to having an engineer (or whoever) teach it?

I am a theoretical and computational person working at the interface of optics and biophysics, and I interact with people possessing every imaginable disciplinary designation (whether self-described or departmentally-imposed). To the extent that my "physicist" label means anything, it means that I bring a different perspective to these problems than my colleagues. Not a better perspective--the field would be poorer if everybody came at it the way I do--just a different one. I like to think that this perspective gives me something unique to bring to bring to classical mechanics, but it could just as easily be that teaching classical mechanics gives me a unique perspective on these problems. Interestingly, I did some work on applying variational calculus to my research in the same year that I taught Lagrangian mechanics...

To expand on it, I like to think of these disciplinary identity issues as something of an "inverted T": We know a few things really, really well, and then we know a bunch of other things to varying degrees of depth. Our research determines what we know really, really well. The other things come from a combination of research (i.e. issues that show up in the work but are not the focus), coursework, teaching, side reading, and just whom you're around (e.g. what you pick up from lunch conversations with colleagues, or attending seminars). My research expertise would make for a reasonable fit in a number of different departments, but my broader body of knowledge pretty clearly marks me as a physicist.

The physics department is paying you, instead of the mechanical engineering department. Clear difference.

Actually, the idea of your spread of background knowledge is a very good one--all the research work I'm doing is mechanical engineering, but I still consider myself a physicist because I understand quantum damnit.

But the more I think about it, the more I think having an engineer teaching basic physics might be a better idea. I think as an engineer you get a much better heuristic understanding of Newtonian physics: the importance of the SHO, what happens in collisions, voltages and current. Much of what I've learned in (honors) freshman physics can be described as "Because of Symmetry", which is beautiful and interesting but also a poor description of what we have to deal with. Leave the "B. (of) S." for when you teach Hamiltonian/Lagrangian.

And on that note, I actually think D.C.'s idea is a very bad one--it will heavily hamper the ability of any physicist undergrad to go into a job that requires a BS in physics level of work, as those tend to be more engineering or geophysics (Prior discussion being…). For all the work I'm doing, I don't need to understand relativistic dilation one iota, but I damn well need to understand what a displacement boundary condition is.

Much of what I’ve learned in (honors) freshman physics can be described as “Because of Symmetry”, which is beautiful and interesting but also a poor description of what we have to deal with.

This is true, but in the real world you also never see conservation laws. There's always some friction that makes it harder to directly observe conservation of mechanical energy. And while that can (at least sometimes) be accounted for in the calculations, what about energy carried away in the form of sound from a collision? That is much harder to treat precisely. And all friction eventually leads to some momentum transfer between the system and the earth (e.g. friction with the ground slows the car, transferring momentum to the earth), so you never even see true momentum transfer in practice.

Despite that, conservation laws are tremendously important. And symmetry is tremendously important, at least as a starting point.

Maybe the best argument for having physicists teach these things is that nobody else will fuss over symmetry and conservation laws as much as we do. I think that's also the worst argument for having us teach these things, of course...

I disagree, I think conservation laws are vital for engineering! How else can we understand heating and cooling and give ourselves useful limits? We already cheat in physics class when we say "some collisions are inelastic" and give a restitution coefficient like we're so smart.

Better idea? Talk about time scales! Here's a physics example: In plasma we often treat a field as going to the lowest energy state with helicity conserved. Helicity itself will slowly decrease, but its the rate limiting step in the whole process. So even in deep physics we need to know when a law is applicable in practice and when to ignore it.

Here's a lovely example tied to my work. Hertz calculated the time of interaction of an elastic sphere rebounding off an infinite, elastic half-space (giant block of metal). It was initially assumed that the collision was elastic for the ball--it must rebound to its original height!

But if you're intelligent, then you know that the net momentum must remain downward just after the collision. So some folks in the 1950's (Hunter, Reed) tried to estimate the work done by the sphere on the half-space, and eventually we were able to account for the energy loss by looking at the waves generated and propagated, fulfilling the momentum conservation requirement. So, paying attention to this helped us understand a fundamental issue in mechanical engineering and contact dynamics.

I disagree, I think conservation laws are vital for engineering! How else can we understand heating and cooling and give ourselves useful limits?

Very true. I'd just add that symmetric problems also provide useful limits, but you discounted those.

If we take it as a given that symmetric problems and conservation laws provide useful limits and are, if nothing else, useful starting points for learning principles, then it's worth asking which disciplines have the greatest number of people who care about symmetric problems and conservation laws for their own sakes.

Not sure I agree. I would offer that chemistry is much more about praxis than physics. Alternately, chemistry has higher informational entropy - less theoretical cohesion. The usual chemistry curriculum is: general; quantitative methods; organic, physical; inorganic;.... After general, none of these courses can be taught by an unspecialist, partly because of techniques and partly because of thoughts.

By Bruce W Fowler (not verified) on 15 Feb 2013 #permalink

It would be interesting to see the ways in which people from different disciplines might teach intro physics. For example, a systems engineer might do something like this:
Assume the mass is an LTI system, the input is thhe applied force and the output is the position of the mass relative to its starting point. Now apply an impulse to the mass (i.e. kick it) and track the trajectory. We now have the full description of the mass for all useful purposes.

I'm not sure I buy it, mostly because I don't think you're comparing like things.

Organic chemistry generally falls 3rd & 4th in the required course sequence (and some, but probably not most, students take additional courses before organic).

Unless I'm misinterpreting the physics courses you are talking about, they are generally the 1st and 2nd courses students take in physics.

If I'm interpreting correctly, that makes the relevant comparison on the chemistry side general chemistry 1 & 2. If that's the comparison, then I don't see things as different on the chemistry side than they are on the physics side. I teach introductory chemistry but I certainly don't do research on it, and neither does anyone else in my department. Yet, just like in physics, somebody has to teach the old stuff, and about half of our department teaches at least one of the introductory courses.

Just to be clear - I actually enjoy teaching 101, but not because the subject matter has anything to do with my research.

It may be interesting to note that the main reason organic has the place it does in the undergrad curriculum (typically sophomore year) is for biology and med students (and the reason we chemists have our jobs is primarily thanks to biology majors and med students). If you asked chemists to design the curriculum for chemists, I'm sure it would focus more on physical chemistry (QM, thermodynamics, and statistical mechanics) and organic would become more of a specialty upper level course (with parts being transferred to the intro courses).

First, you really should sit down and talk to your local engineers about what they want their students to learn from you and why they want physics taught by physicists and not engineers. I have found such discussions enlightening.

Second, since teaching in a CC has me co-located with people teaching first year chemistry and calculus, I'd have to say you have a fairly weak understanding of what is taught in freshman chemistry vis-a-vis what any chemist does in research. I'd hope that your college doesn't face the educational challenges that show up at a large R1 or a CC, but talk to some chemists and find out. Can their students solve a word problem? Can they use algebra and PV=nRT, or do they have to memorize equations for each special case? It can get pretty gnarly outside of elite colleges where most students come in having had a decent HS chemistry class.

I'd say the division between orgo and everything else is for the simple reason that only organic chemists would dare to teach it, and orgo is only a bed of roses if you remember how many grade-grubbing pre-med thorns are in there!

By CCPhysicist (not verified) on 18 Feb 2013 #permalink

BTW, it's very much more so in Statistics. It'd take two years of graduate work to get to anything on which the professors are doing research on.