HN Explorer

Ask HN: Best way to learn modern Physics?

10 points by mudrockbestgirl ↗ HN
I've recently become interested in modern Physics. I want to understand quantum mechanics, space-time, quantum gravity, and what we understand about the nature of reality. I have a decent mathematical background (PhD in CS) but almost no Physics knowledge. I took a few Physics classes in college, but all I did was memorize a few equations and pattern-match them to problem in the exams. I remember nothing.

I could go the traditional route and work through standard Physics books, e.g. something like the Feynman lectures. But if I want to arrive at modern Physics, is it even a good idea to spend many months working through "Newtonian" stuff that may not be relevant? How much of the standard Physics curriculum (Mechanics, Electricity and Magnetism, etc) is a prerequisite for understanding modern Physics? You can see I'm quite clueless.

How do I best go about this? Do you have any resource you would recommend?

EDIT: I took a look at SICM [0] because it has a CS angle which I thought may help, but it's above my level. I don't understand what many of the Physics terms means and I don't have the intuition to follow the explanations.

[0] https://groups.csail.mit.edu/mac/users/gjs/6946/sicm-html/book.html

8 comments

[ 3.6 ms ] story [ 35.6 ms ] thread
jmopp
You do want that basis of classical mechanics, because one of the things it teaches you is how to observe a phenomenon and think about the interactions taking place, building up that intuition you're lacking. Physics requires more than just memorising formulas and pattern-matching. It requires you to properly think about what's going on in the interaction you're looking at.
exmadscientist
You can't really skip anything. Concepts and techniques recur, over and over again. Half of dealing with the "advanced" stuff is just taking its leading order term and pretending it's classical. Quantum electrodynamics is fun for explaining atomic physics and below, but for everything bigger than a molecule it's useless and classical E&M rules the day.

So... yeah, you've got to start at the beginning. Or at least, know what you're passing by.

FiatLuxDave
Everyone learns differently, so you are going to get a lot of different types of suggestions. I learned physics through a mix of the standard academic route and some autodidactic explorations, so I'm not particularly married to the idea of learning things 'in the proper order'. On the other hand, I've experienced what it is like to learn a part of something but to have a hole in your knowledge because you learned it out of the expected order. Filling those holes is helpful, but just because you skipped something doesn't mean you can't learn it later.

I would suggest to pick up a used copy of Modern Physics by Serway Moses and Moyer. You should be able to find an old edition for about $20. Read through it once first, don't try to work the problems, just absorb the concepts. This will give you a basic map of the idea space, so you know what you'll need to learn. Then find a part which you find fascinating and try some of the problems. You'll likely find you'll be missing some tools and then you'll need to go back and learn some things, but at least you will know what you need it for, and your interest will fuel your ability to work through the problems.

SaberTail
You need those cores of mechanics, E&M, and QM. They're so important that most physics majors take them twice, first at an introductory level and then at a more detailed level.

It's only a slight exaggeration to say that the simple harmonic oscillator is the most important model in physics. You learn it in freshman mechanics, and you still need it to understand quantum field theory.

The hard part is that to really understand physics, you need to work through problems. Reading or lectures alone won't do it. You'll need to learn to apply the concepts from what you're learning. If you are following textbooks, definitely work through the problems.

(my credentials for this advice are that I have a PhD in physics)

mudrockbestgirl
Thanks. Your advice is what I needed to hear, though I have have to admit it is not exactly what I wanted to hear. Given my background, I wished there was a shortcut, a way to skip the repetitive and boring work of solving lots and lots of undergraduate physics problems and working through the same material from multiple angles. But after doing some research and reading other advice similar to yours I came to the conclusion that no such thing exists.

If I truly want to understand modern Physics, I am looking at at least 4-5 years of serious study that includes solving lots of textbook problems and repetition.

exmadscientist
Oh, there are plenty of shortcuts. They've just got prices of their own. Problem sets are actually only the second best way to learn the material. The very best way is experiments, which is why any half-decent course comes with a lab section.

Unfortunately, picking problems and setting up experiments is not easy to do. You should actually find that if things have been set up well and you are learning quickly, there is not a lot of repetition. You should be either building on what you already know or revealing another facet of it. You should not be doing things over very often, though some things are very unintuitive (rotating coordinate frames, anyone...) and some repetition there usually helps. Self-directed learning allows for good tuning there and even good delaying... if you are well-regulated. That, of course, is always the hard part with self-directed learning.

The other problem you will face is that your math skills are probably not what you think they are. CS is a branch of discrete mathematics. Physics is the backbone underlying continuous, analytic mathematics... with a bit of higher algebra (symmetries! symmetries everywhere!) sprinkled on top. These domains are different, and it takes either some effort or very great skill to be fluent in both. But physics drove the growth of this type of mathematics, and so chances are good you will find things natural to pick up as you go. (Except for vector fields. Not coincidentally, they came later and were retrofitted onto the physics. They do work and do make things easier, but they're not always intuitive.)

I don't think you're looking at 4–5 years to learn this stuff, unless you are seeking particularly deep understanding and insights, but you're not wrong that it will be some work.