# Why light go straight?

1. Jan 29, 2005

### scilover89

In normal condition, light always go straight. But since photon move randomly, why must they go straight?

2. Jan 29, 2005

### Gokul43201

Staff Emeritus
What do you mean by this ?

3. Jan 30, 2005

### Chronos

Straight with respect to what? Photons usually do take random paths, but they travel on geodesic paths once they chose a course. After that, they appear to travel along the same path, until disturbed.

4. Jan 30, 2005

### Gokul43201

Staff Emeritus
I think the OP may be confusing the random choice of a direction with the randomness that is associated with Brownian motion or the drunken walk.

5. Jan 30, 2005

### misogynisticfeminist

I think what the thread starter wanted is actually said in one of Feynman's books. In classical optics, light is seen as travelling in straight lines, but that is shows as an approximation of what really happens. Using the principle of least action, light does indeed travel in a straight line since that's the easiest path to travel.

But according to QM, a photon can take several crazy trajectories, but if looking at the probability cloud of a photon, its mostly concentrated in a straight line. So in the context of QM, light doesn't really travel in straight lines, it just MOST PROBABLY would travel in a straight line.

Am i right on this one?

6. Jan 30, 2005

### Newton

^ at what speed does it travel on those other paths?

7. Jan 30, 2005

### DB

"c" 299 792 458 m/s

8. Jan 30, 2005

### caribou

A photon in Feynman's sum-over-histories has countless paths faster and slower than the speed-of-light that combine to give the speed-of-light path.

We have to be careful not to think of a particle as having anything like a definite position and velocity at the same time, though, but we can describe it as series of definite positions at all times so as to make up these paths for the sum-over-histories, I think.

Last edited: Jan 30, 2005
9. Jan 30, 2005

### DB

Really? I did not know this. Could you elaborate more on this or maybe suply me with a link. If this topic has a name I'll google it.

10. Jan 30, 2005

### caribou

Actually, the "slower than light" bit I mention might not be right for photons. I remember now that was something I meant to check to see if it was a possibility. Oops.

I know that faster than light and even backwards in time histories are correct and appear in a sum-over-histories, though.

Last edited: Jan 30, 2005
11. Jan 30, 2005

### caribou

Professor Feynman to the rescue! I was right about the slower than light histories for photons in a sum-over-histories. That's a relief.

To quote Richard Feynman "...there is also an amplitude for light to go faster (or slower) than the conventional speed of light. You found out in the last lecture that light doesn't go only in straight lines; now, you find out that it doesn't go only at the speed of light! It may surprise you that there is an amplitude for a photon to go at speeds faster or slower than the conventional speed, c." - Chapter 3, page 89 of Richard Feynman's book "QED".

Got that from:

http://en.wikipedia.org/wiki/Faster_than_light

Last edited: Jan 30, 2005
12. Jan 31, 2005

### Gonzolo

"Momentum" is the explanation that usually serves my purpose.

13. Jan 31, 2005

### masudr

Photons do not move randomly. QM is not just all about randomness - there are very strict rules about how objects can behave. In order to decide the probability that some system will start in classical state A and end in classical state B, we do the Feynman cookbooking (i.e. $$e^{\frac{i}{\hbar}S}$$ where S is the classical action of that history, and then sum this value for every possible history to get the probability (to within some proportionalirt factor) that the system will truly go from state A to state B). So it's not all random, there are precise rules. It was designed so that for macroscopic systems, the outcome is exactly the same as the classical outcome.

Share this great discussion with others via Reddit, Google+, Twitter, or Facebook